fix: system title (#3522)

* split tokens into input and output (#3477)

* split tokens into input and output

* query extension & tool call & question guide

* fix

* perf: input and output tokens

* perf: tool call if else

* perf: remove code

* fix: extract usage count

* fix: qa usage count

---------

Co-authored-by: heheer <heheer@sealos.io>

README.md

@@ -58,9 +58,9 @@ https://github.com/labring/FastGPT/assets/15308462/7d3a38df-eb0e-4388-9250-2409b
    - [x] 多库复用，混用
    - [x] chunk 记录修改和删除
    - [x] 支持手动输入，直接分段，QA 拆分导入
-   - [x] 支持 txt，md，html，pdf，docx，pptx，csv，xlsx (有需要更多可 PR file loader)
-   - [x] 支持 url 读取、CSV 批量导入
+   - [x] 支持 txt，md，html，pdf，docx，pptx，csv，xlsx (有需要更多可 PR file loader)，支持 url 读取、CSV 批量导入
    - [x] 混合检索 & 重排
+   - [x] API 知识库
    - [ ] 自定义文件读取服务
    - [ ] 自定义分块服务
   
@@ -69,7 +69,7 @@ https://github.com/labring/FastGPT/assets/15308462/7d3a38df-eb0e-4388-9250-2409b
    - [x] 对话时反馈引用并可修改与删除
    - [x] 完整上下文呈现
    - [x] 完整模块中间值呈现
-   - [x] 高级编排 DeBug 模式
+   - [ ] 高级编排 DeBug 模式
   
 `4` OpenAPI 接口
    - [x] completions 接口 (chat 模式对齐 GPT 接口)
@@ -104,7 +104,7 @@ https://github.com/labring/FastGPT/assets/15308462/7d3a38df-eb0e-4388-9250-2409b
 * [快速开始本地开发](https://doc.tryfastgpt.ai/docs/development/intro/)
 * [部署 FastGPT](https://doc.tryfastgpt.ai/docs/development/sealos/)
 * [系统配置文件说明](https://doc.tryfastgpt.ai/docs/development/configuration/)
-* [多模型配置](https://doc.tryfastgpt.ai/docs/development/one-api/)
+* [多模型配置方案](https://doc.tryfastgpt.ai/docs/development/modelconfig/one-api/)
 * [版本更新/升级介绍](https://doc.tryfastgpt.ai/docs/development/upgrading/)
 * [OpenAPI API 文档](https://doc.tryfastgpt.ai/docs/development/openapi/)
 * [知识库结构详解](https://doc.tryfastgpt.ai/docs/guide/knowledge_base/rag/)
@@ -127,7 +127,6 @@ https://github.com/labring/FastGPT/assets/15308462/7d3a38df-eb0e-4388-9250-2409b
 
 我们正在寻找志同道合的小伙伴，加速 FastGPT 的发展。你可以通过 [FastGPT 2025 招聘](https://fael3z0zfze.feishu.cn/wiki/P7FOwEmPziVcaYkvVaacnVX1nvg)了解 FastGPT 的招聘信息。
 
-
 ## 💪 相关项目
 
 - [Laf：3 分钟快速接入三方应用](https://github.com/labring/laf)
@@ -139,19 +138,21 @@ https://github.com/labring/FastGPT/assets/15308462/7d3a38df-eb0e-4388-9250-2409b
     <img src="https://img.shields.io/badge/-返回顶部-7d09f1.svg" alt="#" align="right">
 </a>
 
-## 👀 其他
 
-- [保姆级 FastGPT 教程](https://www.bilibili.com/video/BV1n34y1A7Bo/?spm_id_from=333.999.0.0)
-- [接入飞书](https://www.bilibili.com/video/BV1Su4y1r7R3/?spm_id_from=333.999.0.0)
-- [接入企微](https://www.bilibili.com/video/BV1Tp4y1n72T/?spm_id_from=333.999.0.0)
+## 🌿 第三方生态
+
+- [COW 个人微信/企微机器人](https://doc.tryfastgpt.ai/docs/use-cases/external-integration/onwechat/)
+- [SiliconCloud (硅基流动) —— 开源模型在线体验平台](https://cloud.siliconflow.cn/i/TR9Ym0c4)
 
 <a href="#readme">
     <img src="https://img.shields.io/badge/-返回顶部-7d09f1.svg" alt="#" align="right">
 </a>
 
-## 🌿 第三方生态
+## 👀 其他
 
-- [COW 个人微信/企微机器人](https://doc.tryfastgpt.ai/docs/use-cases/external-integration/onwechat/)
+- [保姆级 FastGPT 教程](https://www.bilibili.com/video/BV1n34y1A7Bo/?spm_id_from=333.999.0.0)
+- [接入飞书](https://www.bilibili.com/video/BV1Su4y1r7R3/?spm_id_from=333.999.0.0)
+- [接入企微](https://www.bilibili.com/video/BV1Tp4y1n72T/?spm_id_from=333.999.0.0)
 
 <a href="#readme">
     <img src="https://img.shields.io/badge/-返回顶部-7d09f1.svg" alt="#" align="right">
@@ -214,4 +215,4 @@ https://github.com/labring/FastGPT/assets/15308462/7d3a38df-eb0e-4388-9250-2409b
 1. 允许作为后台服务直接商用，但不允许提供 SaaS 服务。
 2. 未经商业授权，任何形式的商用服务均需保留相关版权信息。
 3. 完整请查看 [FastGPT Open Source License](./LICENSE)
-4. 联系方式：Dennis@sealos.io，[点击查看商业版定价策略](https://doc.tryfastgpt.ai/docs/commercial)
+4. 联系方式：Dennis@sealos.io，[点击查看商业版定价策略](https://doc.tryfastgpt.ai/docs/shopping_cart/intro/)

docSite/content/zh-cn/docs/development/configuration.md

@@ -4,7 +4,7 @@ description: 'FastGPT 配置参数介绍'
 icon: 'settings'
 draft: false
 toc: true
-weight: 708
+weight: 707
 ---
 
 由于环境变量不利于配置复杂的内容，新版 FastGPT 采用了 ConfigMap 的形式挂载配置文件，你可以在 `projects/app/data/config.json` 看到默认的配置文件。可以参考 [docker-compose 快速部署](/docs/development/docker/) 来挂载配置文件。
@@ -97,7 +97,9 @@ weight: 708
       "customExtractPrompt": "",
       "defaultSystemChatPrompt": "",
       "defaultConfig": {
-        "temperature": 1
+        "temperature": 1,
+        "max_tokens": null,
+        "stream": false
       }
     },
     {
@@ -122,7 +124,9 @@ weight: 708
       "customExtractPrompt": "",
       "defaultSystemChatPrompt": "",
       "defaultConfig": {
-        "temperature": 1
+         "temperature": 1,
+        "max_tokens": null,
+        "stream": false
       }
     }
   ],
@@ -164,6 +168,7 @@ weight: 708
   "reRankModels": [],
   "audioSpeechModels": [
     {
+      "provider": "OpenAI",
       "model": "tts-1",
       "name": "OpenAI TTS1",
       "charsPointsPrice": 0,
@@ -178,6 +183,7 @@ weight: 708
     }
   ],
   "whisperModel": {
+    "provider": "OpenAI",
     "model": "whisper-1",
     "name": "Whisper1",
     "charsPointsPrice": 0
@@ -185,7 +191,7 @@ weight: 708
 }
 ```
 
-## 模型提供商
+##  内置的模型提供商ID
 
 为了方便模型分类展示，FastGPT 内置了部分模型提供商的名字和 Logo。如果你期望补充提供商，可[提交 Issue](https://github.com/labring/FastGPT/issues)，并提供几个信息：
 
@@ -197,7 +203,9 @@ weight: 708
 - OpenAI
 - Claude
 - Gemini
+- Meta
 - MistralAI
+- AliCloud - 阿里云
 - Qwen - 通义千问
 - Doubao - 豆包
 - ChatGLM - 智谱
@@ -209,13 +217,40 @@ weight: 708
 - Baichuan - 百川
 - Yi - 零一万物
 - Ernie - 文心一言
+- StepFun - 阶跃星辰
 - Ollama
+- BAAI - 智源研究院
+- FishAudio
 - Other - 其他
 
 
-## 特殊模型
+## ReRank 模型接入
 
-### ReRank 接入(私有部署)
+由于 OneAPI 不支持 Rerank 模型，所以需要单独配置接入，这里
+
+
+### 使用硅基流动的在线模型
+
+有免费的 `bge-reranker-v2-m3` 模型可以使用。
+
+1. [点击注册硅基流动账号](https://cloud.siliconflow.cn/i/TR9Ym0c4)
+2. 进入控制台，获取 API key: https://cloud.siliconflow.cn/account/ak
+3. 修改 FastGPT 配置文件
+
+```json
+{
+    "reRankModels": [
+        {
+            "model": "BAAI/bge-reranker-v2-m3", // 这里的model需要对应 siliconflow 的模型名
+            "name": "BAAI/bge-reranker-v2-m3",
+            "requestUrl": "https://api.siliconflow.cn/v1/rerank",
+            "requestAuth": "siliconflow 上申请的 key"
+        }
+    ]
+}
+```
+
+### 私有部署模型
 
 请使用 4.6.6-alpha 以上版本，配置文件中的 `reRankModels` 为重排模型，虽然是数组，不过目前仅有第1个生效。
 
@@ -236,44 +271,3 @@ weight: 708
     ]
 }
 ```
-
-### ReRank 接入（硅基流动）
-
-有免费的 `bge-reranker-v2-m3` 模型可以使用。
-
-1. 注册硅基流动账号: https://siliconflow.cn/
-2. 进入控制台，获取 API key: https://cloud.siliconflow.cn/account/ak
-3. 修改 FastGPT 配置文件
-
-```json
-{
-    "reRankModels": [
-        {
-            "model": "BAAI/bge-reranker-v2-m3", // 这里的model需要对应 siliconflow 的模型名
-            "name": "BAAI/bge-reranker-v2-m3",
-            "requestUrl": "https://api.siliconflow.cn/v1/rerank",
-            "requestAuth": "siliconflow 上申请的 key"
-        }
-    ]
-}
-```
-
-### ReRank 接入（Cohere）
-
-这个重排模型对中文不是很好，不如 bge 的好用。
-
-1. 申请 Cohere 官方 Key: https://dashboard.cohere.com/api-keys
-2. 修改 FastGPT 配置文件
-
-```json
-{
-    "reRankModels": [
-        {
-            "model": "rerank-multilingual-v2.0", // 这里的model需要对应 cohere 的模型名
-            "name": "rerank-multilingual-v2.0",
-            "requestUrl": "https://api.cohere.ai/v1/rerank",
-            "requestAuth": "Coherer上申请的key"
-        }
-    ]
-}
-```

docSite/content/zh-cn/docs/development/custom-models/chatglm2-m3e.md

@@ -23,7 +23,7 @@ FastGPT 默认使用了 OpenAI 的 LLM 模型和向量模型，如果想要私
 也可以通过环境变量引入：sk-key。有关docker环境变量引入的方法请自寻教程，此处不再赘述。
 ```
 
-## 接入 [One API](/docs/development/one-api/)
+## 接入 [One API](/docs/development/modelconfig/one-api/)
 
 为 chatglm2 和 m3e-large 各添加一个渠道，参数如下：
 

docSite/content/zh-cn/docs/development/custom-models/xinference.md

@@ -102,7 +102,7 @@ xinference launch -n qwen-chat -s 14 -f pytorch
 
 ## 将本地模型接入 One API
 
-One API 的部署和接入请参考[这里](/docs/development/one-api/)。
+One API 的部署和接入请参考[这里](/docs/development/modelconfig/one-api/)。
 
 为 qwen1.5-chat 添加一个渠道，这里的 Base URL 需要填 Xinference 服务的端点，并且注册 qwen-chat (模型的 UID) 。
 

docSite/content/zh-cn/docs/development/docker.md

@@ -192,7 +192,7 @@ docker restart oneapi
 
 可以通过`ip:3001`访问OneAPI，默认账号为`root`密码为`123456`。
 
-在OneApi中添加合适的AI模型渠道。[点击查看相关教程](/docs/development/one-api/)
+在OneApi中添加合适的AI模型渠道。[点击查看相关教程](/docs/development/modelconfig/one-api/)
 
 ### 5. 访问 FastGPT
 

docSite/content/zh-cn/docs/development/faq.md

@@ -1,5 +1,5 @@
 ---
-weight: 749
+weight: 740
 title: "私有部署常见问题"
 description: "FastGPT 私有部署常见问题"
 icon: upgrade

docSite/content/zh-cn/docs/development/modelConfig/_index.md

@@ -0,0 +1,8 @@
+---
+weight: 745
+title: '模型配置方案'
+description: '本模型配置方案'
+icon: 'code_blocks'
+draft: false
+images: []
+---

docSite/content/zh-cn/docs/development/modelConfig/one-api.md

@@ -0,0 +1,189 @@
+---
+title: '通过 OneAPI 接入模型'
+description: '通过 OneAPI 接入模型'
+icon: 'api'
+draft: false
+toc: true
+weight: 745
+---
+
+FastGPT 目前采用模型分离的部署方案，FastGPT 中只兼容 OpenAI 的模型规范（OpenAI 不存在的模型采用一个较为通用的规范），并通过 [One API](https://github.com/songquanpeng/one-api) 来实现对不同模型接口的统一。
+
+[One API](https://github.com/songquanpeng/one-api) 是一个 OpenAI 接口管理 & 分发系统，可以通过标准的 OpenAI API 格式访问所有的大模型，开箱即用。
+
+
+## FastGPT 与 One API 关系
+
+可以把 One API 当做一个网关，FastGPT 与 One API 关系：
+
+![](/imgs/sealos-fastgpt.webp)
+
+## 部署
+
+### Docker 版本
+
+`docker-compose.yml` 文件已加入了 OneAPI 配置，可直接使用。默认暴露在 3001 端口。
+
+### Sealos 版本
+
+* 北京区: [点击部署 OneAPI](https://hzh.sealos.run/?openapp=system-template%3FtemplateName%3Done-api)
+* 新加坡区(可用 GPT) [点击部署 OneAPI](https://cloud.sealos.io/?openapp=system-template%3FtemplateName%3Done-api)
+
+![alt text](/imgs/image-59.png)
+
+部署完后，可以打开 OneAPI 访问链接，进行下一步操作。
+
+## OneAPI 基础教程
+
+### 概念
+
+1. 渠道：
+   1. OneApi 中一个渠道对应一个 `Api Key`，这个 `Api Key` 可以是GPT、微软、ChatGLM、文心一言的。一个`Api Key`通常可以调用同一个厂商的多个模型。
+   2. One API 会根据请求传入的`模型`来决定使用哪一个`渠道`，如果一个模型对应了多个`渠道`，则会随机调用。
+2. 令牌：访问 One API 所需的凭证，只需要这`1`个凭证即可访问`One API`上配置的模型。因此`FastGPT`中，只需要配置`One API`的`baseurl`和`令牌`即可。令牌不要设置任何的模型范围权限，否则容易报错。
+
+![alt text](/imgs/image-60.png)
+
+### 大致工作流程
+
+1. 客户端请求 One API
+2. 根据请求中的 `model` 参数，匹配对应的渠道（根据渠道里的模型进行匹配，必须完全一致）。如果匹配到多个渠道，则随机选择一个（同优先级）。
+3. One API 向真正的地址发出请求。
+4. One API 将结果返回给客户端。
+
+### 1. 登录 One API
+
+![step5](/imgs/oneapi-step5.png)
+
+### 2. 创建渠道
+
+在 One API 中添加对应渠道，直接点击 【添加基础模型】，不要遗漏了向量模型（Embedding）
+
+![step6](/imgs/oneapi-step6.png)
+
+### 3. 创建令牌
+
+| | |
+| --- | --- |
+| ![step7](/imgs/oneapi-step7.png) | ![alt text](/imgs/image-61.png) |
+
+### 4. 修改账号余额
+
+One API 默认 root 用户只有 200刀，可以自行修改编辑。
+
+![alt text](/imgs/image-62.png)
+
+### 5. 修改 FastGPT 的环境变量
+
+有了 One API 令牌后，FastGPT 可以通过修改 `baseurl` 和 `key` 去请求到 One API，再由 One API 去请求不同的模型。修改下面两个环境变量：
+
+```bash
+# 务必写上 v1。如果在同一个网络内，可改成内网地址。
+OPENAI_BASE_URL=https://xxxx.cloud.sealos.io/v1
+# 下面的 key 是由 One API 提供的令牌
+CHAT_API_KEY=sk-xxxxxx
+```
+
+## 接入其他模型
+
+**以添加文心一言为例:**
+
+### 1. OneAPI 新增模型渠道
+
+类型选择百度文心千帆。
+
+![](/imgs/oneapi-demo1.png)
+
+### 2. 修改 FastGPT 配置文件
+
+可以在 `/projects/app/src/data/config.json` 里找到配置文件（本地开发需要复制成 config.local.json）,按下面内容修改配置文件，最新/更具体的配置说明，可查看[FastGPT 配置文件说明](/docs/development/configuration)。
+
+配置模型关键点在于`model` 需要与 OneAPI 渠道中的模型一致。
+
+```json
+{
+  "llmModels": [ // 语言模型配置
+    {
+      "model": "ERNIE-Bot", // 这里的模型需要对应 One API 的模型
+      "name": "文心一言", // 对外展示的名称
+      "avatar": "/imgs/model/openai.svg", // 模型的logo
+      "maxContext": 16000, // 最大上下文
+      "maxResponse": 4000, // 最大回复
+      "quoteMaxToken": 13000, // 最大引用内容
+      "maxTemperature": 1.2, // 最大温度
+      "charsPointsPrice": 0, 
+      "censor": false,
+      "vision": false, // 是否支持图片输入
+      "datasetProcess": true, // 是否设置为知识库处理模型
+      "usedInClassify": true, // 是否用于问题分类
+      "usedInExtractFields": true, // 是否用于字段提取
+      "usedInToolCall": true, // 是否用于工具调用
+      "usedInQueryExtension": true, // 是否用于问题优化
+      "toolChoice": true, // 是否支持工具选择
+      "functionCall": false, // 是否支持函数调用
+      "customCQPrompt": "", // 自定义文本分类提示词（不支持工具和函数调用的模型
+      "customExtractPrompt": "", // 自定义内容提取提示词
+      "defaultSystemChatPrompt": "", // 对话默认携带的系统提示词
+      "defaultConfig":{}  // 请求API时，挟带一些默认配置（比如 GLM4 的 top_p）
+    }
+  ],
+  "vectorModels": [ // 向量模型配置
+    {
+      "model": "text-embedding-ada-002",
+      "name": "Embedding-2",
+      "avatar": "/imgs/model/openai.svg",
+      "charsPointsPrice": 0,
+      "defaultToken": 700,
+      "maxToken": 3000,
+      "weight": 100
+    },
+  ]
+}
+```
+
+### 3. 重启 FastGPT
+
+**Docker 版本**
+
+```bash
+docker-compose down
+docker-compose up -d
+```
+
+**Sealos 版本**
+
+直接找到 FastGPT 服务，点击重启即可。
+
+
+## 其他服务商接入参考
+
+这章介绍一些提供商接入 OneAPI 的教程，配置后不要忘记修改 FastGPT 配置文件。
+
+### 阿里通义千问
+
+千问目前已经兼容 GPT 格式，可以直接选择 OpenAI 类型来接入即可。如下图，选择类型为`OpenAI`，代理填写阿里云的代理地址。
+
+目前可以直接使用阿里云的语言模型和 `text-embedding-v3` 向量模型（实测已经归一化，可直接使用）
+
+![alt text](/imgs/image-63.png)
+
+### 硅基流动 —— 开源模型大合集
+
+[硅基流动](https://cloud.siliconflow.cn/i/TR9Ym0c4) 是一个专门提供开源模型调用平台，并拥有自己的加速引擎。模型覆盖面广，非常适合低成本来测试开源模型。接入教程：
+
+1. [点击注册硅基流动账号](https://cloud.siliconflow.cn/i/TR9Ym0c4)
+2. 进入控制台，获取 API key: https://cloud.siliconflow.cn/account/ak
+3. 新增 OneAPI 渠道，选择`OpenAI`类型，代理填写：`https://api.siliconflow.cn`，密钥是第二步创建的密钥。
+
+![alt text](/imgs/image-64.png)
+
+由于 OneAPI 未内置 硅基流动 的模型名，可以通过自定义模型名称来填入，下面是获取模型名称的教程：
+
+1. 打开[硅基流动模型列表](https://siliconflow.cn/zh-cn/models)
+2. 单击模型后，会打开模型详情。
+3. 复制模型名到 OneAPI 中。
+
+| | | |
+| --- | --- | --- |
+| ![alt text](/imgs/image-65.png) | ![alt text](/imgs/image-66.png)| ![alt text](/imgs/image-67.png) |
+

docSite/content/zh-cn/docs/development/modelConfig/siliconCloud.md

@@ -0,0 +1,220 @@
+---
+title: '通过 SiliconCloud 体验开源模型'
+description: '通过 SiliconCloud 体验开源模型'
+icon: 'api'
+draft: false
+toc: true
+weight: 746
+---
+
+[SiliconCloud(硅基流动)](https://cloud.siliconflow.cn/i/TR9Ym0c4) 是一个以提供开源模型调用为主的平台，并拥有自己的加速引擎。帮助用户低成本、快速的进行开源模型的测试和使用。实际体验下来，他们家模型的速度和稳定性都非常不错，并且种类丰富，覆盖语言、向量、重排、TTS、STT、绘图、视频生成模型，可以满足 FastGPT 中所有模型需求。
+
+如果你想部分模型使用 SiliconCloud 的模型，可额外参考[OneAPI接入硅基流动](/docs/development/modelconfig/one-api/#硅基流动--开源模型大合集)。
+
+本文会介绍完全使用 SiliconCloud 模型来部署 FastGPT 的方案。
+
+
+## 1. 注册 SiliconCloud 账号
+
+1. [点击注册硅基流动账号](https://cloud.siliconflow.cn/i/TR9Ym0c4)
+2. 进入控制台，获取 API key: https://cloud.siliconflow.cn/account/ak
+
+## 2. 修改 FastGPT 环境变量
+
+```bash
+OPENAI_BASE_URL=https://api.siliconflow.cn/v1
+# 填写 SiliconCloud 控制台提供的 Api Key
+CHAT_API_KEY=sk-xxxxxx
+```
+
+## 3. 修改 FastGPT 配置文件
+
+我们选取 SiliconCloud 中的模型作为 FastGPT 配置。这里配置了 `Qwen2.5 72b` 的纯语言和视觉模型；选择 `bge-m3` 作为向量模型；选择 `bge-reranker-v2-m3` 作为重排模型。选择 `fish-speech-1.5` 作为语音模型；选择 `SenseVoiceSmall` 作为语音输入模型。
+
+注意：ReRank 模型仍需配置一次 Api Key
+
+```json
+{
+    "llmModels": [
+    {
+      "provider": "Other", // 模型提供商，主要用于分类展示，目前已经内置提供商包括：https://github.com/labring/FastGPT/blob/main/packages/global/core/ai/provider.ts, 可 pr 提供新的提供商，或直接填写 Other
+      "model": "Qwen/Qwen2.5-72B-Instruct", // 模型名(对应OneAPI中渠道的模型名)
+      "name": "Qwen2.5-72B-Instruct", // 模型别名
+      "maxContext": 32000, // 最大上下文
+      "maxResponse": 4000, // 最大回复
+      "quoteMaxToken": 30000, // 最大引用内容
+      "maxTemperature": 1, // 最大温度
+      "charsPointsPrice": 0, // n积分/1k token（商业版）
+      "censor": false, // 是否开启敏感校验（商业版）
+      "vision": false, // 是否支持图片输入
+      "datasetProcess": true, // 是否设置为文本理解模型（QA），务必保证至少有一个为true，否则知识库会报错
+      "usedInClassify": true, // 是否用于问题分类（务必保证至少有一个为true）
+      "usedInExtractFields": true, // 是否用于内容提取（务必保证至少有一个为true）
+      "usedInToolCall": true, // 是否用于工具调用（务必保证至少有一个为true）
+      "usedInQueryExtension": true, // 是否用于问题优化（务必保证至少有一个为true）
+      "toolChoice": true, // 是否支持工具选择（分类，内容提取，工具调用会用到。）
+      "functionCall": false, // 是否支持函数调用（分类，内容提取，工具调用会用到。会优先使用 toolChoice，如果为false，则使用 functionCall，如果仍为 false，则使用提示词模式）
+      "customCQPrompt": "", // 自定义文本分类提示词（不支持工具和函数调用的模型
+      "customExtractPrompt": "", // 自定义内容提取提示词
+      "defaultSystemChatPrompt": "", // 对话默认携带的系统提示词
+      "defaultConfig": {}, // 请求API时，挟带一些默认配置（比如 GLM4 的 top_p）
+      "fieldMap": {} // 字段映射（o1 模型需要把 max_tokens 映射为 max_completion_tokens）
+    },
+    {
+      "provider": "Other",
+      "model": "Qwen/Qwen2-VL-72B-Instruct",
+      "name": "Qwen2-VL-72B-Instruct",
+      "maxContext": 32000,
+      "maxResponse": 4000,
+      "quoteMaxToken": 30000,
+      "maxTemperature": 1,
+      "charsPointsPrice": 0,
+      "censor": false,
+      "vision": true,
+      "datasetProcess": false,
+      "usedInClassify": false,
+      "usedInExtractFields": false,
+      "usedInToolCall": false,
+      "usedInQueryExtension": false,
+      "toolChoice": false,
+      "functionCall": false,
+      "customCQPrompt": "",
+      "customExtractPrompt": "",
+      "defaultSystemChatPrompt": "",
+      "defaultConfig": {}
+    }
+  ],
+  "vectorModels": [
+    {
+      "provider": "Other",
+      "model": "Pro/BAAI/bge-m3",
+      "name": "Pro/BAAI/bge-m3",
+      "charsPointsPrice": 0,
+      "defaultToken": 512,
+      "maxToken": 5000,
+      "weight": 100
+    }
+  ],
+  "reRankModels": [
+    {
+        "model": "BAAI/bge-reranker-v2-m3", // 这里的model需要对应 siliconflow 的模型名
+        "name": "BAAI/bge-reranker-v2-m3",
+        "requestUrl": "https://api.siliconflow.cn/v1/rerank",
+        "requestAuth": "siliconflow 上申请的 key"
+    }
+  ],
+  "audioSpeechModels": [
+    {
+        "model": "fishaudio/fish-speech-1.5",
+        "name": "fish-speech-1.5",
+        "voices": [
+            {
+                "label": "fish-alex",
+                "value": "fishaudio/fish-speech-1.5:alex",
+                "bufferId": "fish-alex"
+            },
+            {
+                "label": "fish-anna",
+                "value": "fishaudio/fish-speech-1.5:anna",
+                "bufferId": "fish-anna"
+            },
+            {
+                "label": "fish-bella",
+                "value": "fishaudio/fish-speech-1.5:bella",
+                "bufferId": "fish-bella"
+            },
+            {
+                "label": "fish-benjamin",
+                "value": "fishaudio/fish-speech-1.5:benjamin",
+                "bufferId": "fish-benjamin"
+            },
+            {
+                "label": "fish-charles",
+                "value": "fishaudio/fish-speech-1.5:charles",
+                "bufferId": "fish-charles"
+            },
+            {
+                "label": "fish-claire",
+                "value": "fishaudio/fish-speech-1.5:claire",
+                "bufferId": "fish-claire"
+            },
+            {
+                "label": "fish-david",
+                "value": "fishaudio/fish-speech-1.5:david",
+                "bufferId": "fish-david"
+            },
+            {
+                "label": "fish-diana",
+                "value": "fishaudio/fish-speech-1.5:diana",
+                "bufferId": "fish-diana"
+            }
+        ]
+    }
+  ],
+  "whisperModel": {
+    "model": "FunAudioLLM/SenseVoiceSmall",
+    "name": "SenseVoiceSmall",
+    "charsPointsPrice": 0
+  }
+}
+```
+
+## 4. 重启 FastGPT
+
+## 5. 体验测试
+
+### 测试对话和图片识别
+
+随便新建一个简易应用，选择对应模型，并开启图片上传后进行测试：
+
+| | |
+| --- | --- |
+| ![alt text](/imgs/image-68.png) | ![alt text](/imgs/image-70.png) |
+
+可以看到，72B 的模型，性能还是非常快的，这要是本地没几个 4090，不说配置环境，输出怕都要 30s 了。
+
+### 测试知识库导入和知识库问答
+
+新建一个知识库（由于只配置了一个向量模型，页面上不会展示向量模型选择）
+
+| | |
+| --- | --- |
+| ![alt text](/imgs/image-72.png) | ![alt text](/imgs/image-71.png) |
+
+导入本地文件，直接选择文件，然后一路下一步即可。79 个索引，大概花了 20s 的时间就完成了。现在我们去测试一下知识库问答。
+
+首先回到我们刚创建的应用，选择知识库，调整一下参数后即可开始对话：
+
+| | | | 
+| --- | --- | --- |
+| ![alt text](/imgs/image-73.png) | ![alt text](/imgs/image-75.png) | ![alt text](/imgs/image-76.png) |
+
+对话完成后，点击底部的引用，可以查看引用详情，同时可以看到具体的检索和重排得分：
+
+| | |
+| --- | --- |
+| ![alt text](/imgs/image-77.png) | ![alt text](/imgs/image-78.png) |
+
+### 测试语音播放
+
+继续在刚刚的应用中，左侧配置中找到语音播放，点击后可以从弹窗中选择语音模型，并进行试听：
+
+![alt text](/imgs/image-79.png)
+
+### 测试语言输入
+
+继续在刚刚的应用中，左侧配置中找到语音输入，点击后可以从弹窗中开启语言输入
+
+![alt text](/imgs/image-80.png)
+
+开启后，对话输入框中，会增加一个话筒的图标，点击可进行语音输入：
+
+| | |
+| --- | --- |
+| ![alt text](/imgs/image-81.png) | ![alt text](/imgs/image-82.png) |
+
+## 总结
+
+如果你想快速的体验开源模型或者快速的使用 FastGPT，不想在不同服务商申请各类 Api Key，那么可以选择 SiliconCloud 的模型先进行快速体验。
+
+如果你决定未来私有化部署模型和 FastGPT，前期可通过 SiliconCloud 进行测试验证，后期再进行硬件采购，减少 POC 时间和成本。

docSite/content/zh-cn/docs/development/one-api.md

@@ -1,179 +0,0 @@
----
-title: '使用 One API 接入 Azure、ChatGLM 和本地模型'
-description: '部署和使用 One API，实现 Azure、ChatGLM 和本地模型的接入。'
-icon: 'api'
-draft: false
-toc: true
-weight: 708
----
-
-* 默认情况下，FastGPT 只配置了 GPT 的模型，如果你需要接入其他模型，需要进行一些额外配置。
-* [One API](https://github.com/songquanpeng/one-api) 是一个 OpenAI 接口管理 & 分发系统，可以通过标准的 OpenAI API 格式访问所有的大模型，开箱即用。
-* FastGPT 可以通过接入 One API 来实现对不同大模型的支持。One API 的部署方法也很简单。
-
-## FastGPT 与 One API 关系
-
-可以把 One API 当做一个网关。
-
-![](/imgs/sealos-fastgpt.webp)
-
-## 部署
-
-### Docker 版本
-
-已加入最新的 `docker-compose.yml` 文件中。
-
-### Sealos - MySQL 版本
-
-MySQL 版本支持多实例，高并发。
-
-直接点击以下按钮即可一键部署 👇
-
-<a href="https://template.cloud.sealos.io/deploy?templateName=one-api" rel="external" target="_blank"><img src="https://cdn.jsdelivr.net/gh/labring-actions/templates@main/Deploy-on-Sealos.svg" alt="Deploy on Sealos"/></a>
-
-部署完后会跳转「应用管理」，数据库在另一个应用「数据库」中。需要等待 1~3 分钟数据库运行后才能访问成功。
-
-### Sealos - SqlLite 版本
-
-SqlLite 版本不支持多实例，适合个人小流量使用，但是价格非常便宜。
-
-**1. [点击打开 Sealos 公有云](https://cloud.sealos.io/)**
-
-**2. 打开 AppLaunchpad(应用管理) 工具**
-
-![step1](/imgs/oneapi-step1.webp)
-
-**3. 点击创建新应用**
-
-**4. 填写对应参数**
-
-镜像：ghcr.io/songquanpeng/one-api:latest
-
-![step2](/imgs/oneapi-step2.png)
-打开外网访问开关后，Sealos 会自动分配一个可访问的地址，不需要自己配置。
-
-![step3](/imgs/oneapi-step3.png)
-填写完参数后，点击右上角部署即可。环境变量：
-
-```
-SESSION_SECRET=SESSION_SECRET
-POLLING_INTERVAL=60
-BATCH_UPDATE_ENABLED=true
-BATCH_UPDATE_INTERVAL=60
-```
-
-## One API 使用教程
-
-### 概念
-
-1. 渠道：
-   1. OneApi 中一个渠道对应一个 `Api Key`，这个 `Api Key` 可以是GPT、微软、ChatGLM、文心一言的。一个`Api Key`通常可以调用同一个厂商的多个模型。
-   2. One API 会根据请求传入的`模型`来决定使用哪一个`Key`，如果一个模型对应了多个`Key`，则会随机调用。
-2. 令牌：访问 One API 所需的凭证，只需要这`1`个凭证即可访问`One API`上配置的模型。因此`FastGPT`中，只需要配置`One API`的`baseurl`和`令牌`即可。
-
-### 大致工作流程
-
-1. 客户端请求 One API
-2. 根据请求中的 `model` 参数，匹配对应的渠道（根据渠道里的模型进行匹配，必须完全一致）。如果匹配到多个渠道，则随机选择一个（同优先级）。
-3. One API 向真正的地址发出请求。
-4. One API 将结果返回给客户端。
-
-### 1. 登录 One API
-
-打开 【One API 应用详情】，找到访问地址：
-![step4](/imgs/oneapi-step4.png)
-
-登录 One API
-![step5](/imgs/oneapi-step5.png)
-
-### 2. 创建渠道和令牌
-
-在 One API 中添加对应渠道，直接点击 【添加基础模型】，不要遗漏了向量模型（Embedding）
-![step6](/imgs/oneapi-step6.png)
-
-创建一个令牌
-![step7](/imgs/oneapi-step7.png)
-
-### 3. 修改账号余额
-
-One API 默认 root 用户只有 200刀，可以自行修改编辑。
-
-### 4. 修改 FastGPT 的环境变量
-
-有了 One API 令牌后，FastGPT 可以通过修改 `baseurl` 和 `key` 去请求到 One API，再由 One API 去请求不同的模型。修改下面两个环境变量：
-
-```bash
-# 下面的地址是 Sealos 提供的，务必写上 v1， 两个项目都在 sealos 部署时候，https://xxxx.cloud.sealos.io 可以改用内网地址
-OPENAI_BASE_URL=https://xxxx.cloud.sealos.io/v1
-# 下面的 key 是由 One API 提供的令牌
-CHAT_API_KEY=sk-xxxxxx
-```
-
-## 接入其他模型
-
-**以添加文心一言为例:**
-
-### 1. One API 添加对应模型渠道
-
-![](/imgs/oneapi-demo1.png)
-
-### 2. 修改 FastGPT 配置文件
-
-可以在 `/projects/app/src/data/config.json` 里找到配置文件（本地开发需要复制成 config.local.json），配置文件中有一项是**对话模型配置**：
-
-```json
-"llmModels": [
-    ...
-    {
-      "model": "ERNIE-Bot", // 这里的模型需要对应 One API 的模型
-      "name": "文心一言", // 对外展示的名称
-      "avatar": "/imgs/model/openai.svg", // 模型的logo
-      "maxContext": 16000, // 最大上下文
-      "maxResponse": 4000, // 最大回复
-      "quoteMaxToken": 13000, // 最大引用内容
-      "maxTemperature": 1.2, // 最大温度
-      "charsPointsPrice": 0, 
-      "censor": false,
-      "vision": false, // 是否支持图片输入
-      "datasetProcess": true, // 是否设置为知识库处理模型
-      "usedInClassify": true, // 是否用于问题分类
-      "usedInExtractFields": true, // 是否用于字段提取
-      "usedInToolCall": true, // 是否用于工具调用
-      "usedInQueryExtension": true, // 是否用于问题优化
-      "toolChoice": true, // 是否支持工具选择
-      "functionCall": false, // 是否支持函数调用
-      "customCQPrompt": "", // 自定义文本分类提示词（不支持工具和函数调用的模型
-      "customExtractPrompt": "", // 自定义内容提取提示词
-      "defaultSystemChatPrompt": "", // 对话默认携带的系统提示词
-      "defaultConfig":{}  // 请求API时，挟带一些默认配置（比如 GLM4 的 top_p）
-    }
-    ...
-],
-```
-
-**添加向量模型:**
-
-```json
-"vectorModels": [
-  ......
-    {
-      "model": "text-embedding-ada-002",
-      "name": "Embedding-2",
-      "avatar": "/imgs/model/openai.svg",
-      "charsPointsPrice": 0,
-      "defaultToken": 700,
-      "maxToken": 3000,
-      "weight": 100
-    },
-  ......
-]
-```
-
-### 3. 重启 FastGPT
-
-```bash
-docker-compose down
-docker-compose up -d
-```
-
-重启 FastGPT 即可在选择文心一言模型进行对话。**添加向量模型也是类似操作，增加到 `vectorModels`里。**

docSite/content/zh-cn/docs/development/openapi/dataset.md

@@ -1424,7 +1424,11 @@ curl --location --request POST 'https://api.fastgpt.in/api/core/dataset/searchTe
     "limit": 5000,
     "similarity": 0,
     "searchMode": "embedding",
-    "usingReRank": false
+    "usingReRank": false,
+
+    "datasetSearchUsingExtensionQuery": true,
+    "datasetSearchExtensionModel": "gpt-4o-mini",
+    "datasetSearchExtensionBg": ""
 }'
 ```
 
@@ -1441,6 +1445,9 @@ curl --location --request POST 'https://api.fastgpt.in/api/core/dataset/searchTe
 - similarity - 最低相关度（0~1，可选）
 - searchMode - 搜索模式：embedding | fullTextRecall | mixedRecall
 - usingReRank - 使用重排
+- datasetSearchUsingExtensionQuery - 使用问题优化
+- datasetSearchExtensionModel - 问题优化模型
+- datasetSearchExtensionBg - 问题优化背景描述
 {{% /alert %}}
 
 {{< /markdownify >}}

docSite/content/zh-cn/docs/development/sealos.md

@@ -15,7 +15,7 @@ weight: 706
 
 FastGPT 使用了 one-api 项目来管理模型池，其可以兼容 OpenAI 、Azure 、国内主流模型和本地模型等。
 
-可参考：[Sealos 快速部署 OneAPI](/docs/development/one-api)
+可参考：[Sealos 快速部署 OneAPI](/docs/development/modelconfig/one-api)
 
 
 ## 一键部署
@@ -163,4 +163,4 @@ SYSTEM_FAVICON 可以是一个网络地址
 
 ### One API 使用
 
-[参考 OneAPI 使用步骤](/docs/development/one-api/)
+[参考 OneAPI 使用步骤](/docs/development/modelconfig/one-api/)

docSite/content/zh-cn/docs/development/upgrading/4816.md

@@ -1,5 +1,5 @@
 ---
-title: 'V4.8.16(进行中)'
+title: 'V4.8.16(更新配置文件)'
 description: 'FastGPT V4.8.16 更新说明'
 icon: 'upgrade'
 draft: false
@@ -11,9 +11,9 @@ weight: 808
 
 ### 1. 更新镜像：
 
-- 更新 fastgpt 镜像 tag: v4.8.16-beta
-- 更新 fastgpt-pro 商业版镜像 tag: v4.8.16-beta
-- Sandbox 镜像 tag: v4.8.16-beta
+- 更新 fastgpt 镜像 tag: v4.8.16
+- 更新 fastgpt-pro 商业版镜像 tag: v4.8.16
+- Sandbox 镜像 tag: v4.8.16
 
 ### 2. 更新配置文件
 
@@ -68,4 +68,6 @@ weight: 808
 16. 修复 - 简易模式转工作流，没有把系统配置项转化。
 17. 修复 - 插件独立运行，变量初始值未赋上。
 18. 修复 - 工作流使用弹窗组件时，关闭弹窗后，有时候会出现页面偏移。
-19. 修复 - 插件调试时，日志未保存插件输入参数。
+19. 修复 - 插件调试时，日志未保存插件输入参数。
+20. 修复 - 部分模板市场模板
+21. 修复 - 设置NEXT_PUBLIC_BASE_URL时，图片文件读取URL不正确

docSite/content/zh-cn/docs/development/upgrading/4817.md

@@ -0,0 +1,52 @@
+---
+title: 'V4.8.17(包含升级脚本)'
+description: 'FastGPT V4.8.17 更新说明'
+icon: 'upgrade'
+draft: false
+toc: true
+weight: 807
+---
+
+## 更新指南
+
+### 1. 更新镜像：
+
+- 更新 fastgpt 镜像 tag: v4.8.17
+- 更新 fastgpt-pro 商业版镜像 tag: v4.8.17
+- Sandbox 镜像无需更新
+
+
+### 2. 运行升级脚本
+
+从任意终端，发起 1 个 HTTP 请求。其中 {{rootkey}} 替换成环境变量里的 `rootkey`；{{host}} 替换成**FastGPT 域名**。
+
+```bash
+curl --location --request POST 'https://{{host}}/api/admin/initv4817' \
+--header 'rootkey: {{rootkey}}' \
+--header 'Content-Type: application/json'
+```
+
+会将用户绑定的 OpenAI 账号移动到团队中。
+
+
+## 调整 completions 接口返回值
+
+/api/v1/chat/completions 接口返回值调整，对话节点、工具节点等使用到模型的节点，将不再返回 `tokens` 字段，改为返回 `inputTokens` 和 `outputTokens` 字段，分别表示输入和输出的 Token 数量。
+
+## 完整更新内容
+
+1. 新增 - 简易模式工具调用支持数组类型插件。
+2. 新增 - 工作流增加异常离开自动保存，避免工作流丢失。
+3. 新增 - LLM 模型参数支持关闭 max_tokens 和 temperature。
+4. 新增 - 商业版支持后台配置模板市场。
+5. 新增 - 商业版支持后台配置自定义工作流变量，用于与业务系统鉴权打通。
+6. 新增 - 搜索测试接口支持问题优化。
+7. 新增 - 工作流中 Input Token 和 Output Token 分开记录展示。并修复部分请求未记录输出 Token 计费问题。
+8. 优化 - Markdown 大小测试，超出 20 万字符不使用 Markdown 组件，避免崩溃。
+9. 优化 - 知识库搜索参数，滑动条支持输入模式，可以更精准的控制。
+10. 优化 - 可用模型展示UI。
+11. 优化 - Mongo 查询语句，增加 virtual 字段。
+12. 修复 - 文件返回接口缺少 Content-Length 头，导致通过非同源文件上传时，阿里 vision 模型无法识别图片。
+13. 修复 - 去除判断器两端字符串隐藏换行符，避免判断器失效。
+14. 修复 - 变量更新节点，手动输入更新内容时候，非字符串类型数据类型无法自动转化。
+15. 修复 - 豆包模型无法工具调用。

docSite/content/zh-cn/docs/guide/course/ai_settings.md

@@ -44,9 +44,9 @@ weight: 104
 
 被放置在上下文数组的最前面，role 为 system，用于引导模型。
 
-### 最大对话轮数（仅简易模式）
+### 记忆轮数（仅简易模式）
 
-可以配置模型支持的最大对话轮数，如果模型的超出上下文，系统会自动截断，尽可能保证不超模型上下文。
+可以配置模型支持的记忆轮数，如果模型的超出上下文，系统会自动截断，尽可能保证不超模型上下文。
 
 所以尽管配置 30 轮对话，实际运行时候，不一定会达到 30 轮。
 

docSite/content/zh-cn/docs/guide/workbench/intro.md

@@ -45,7 +45,7 @@ FastGPT 从 V4 版本开始采用新的交互方式来构建 AI 应用。使用
 
 ## 重点 - 工作流是如何运行的
 
-FastGPT的工作流从【流程开始】节点开始执行，可以理解为从用户输入问题开始，没有**固定的出口**，是以节点运行结束作为出口，如果在一个轮调用中，所有节点都不再允许，则工作流结束。
+FastGPT的工作流从【流程开始】节点开始执行，可以理解为从用户输入问题开始，没有**固定的出口**，是以节点运行结束作为出口，如果在一个轮调用中，所有节点都不再运行，则工作流结束。
 
 下面我们来看下，工作流是如何运行的，以及每个节点何时被触发执行。
 

docSite/content/zh-cn/docs/guide/workbench/workflow/ai_chat.md

@@ -19,7 +19,7 @@ weight: 232
 
 ## AI模型
 
-可以通过 [config.json](/docs/development/configuration/) 配置可选的对话模型，通过 [one-api](/docs/development/one-api/) 来实现多模型接入。
+可以通过 [config.json](/docs/development/configuration/) 配置可选的对话模型，通过 [one-api](/docs/development/modelconfig/one-api/) 来实现多模型接入。
 
 点击AI模型后，可以配置模型的相关参数。
 

docSite/content/zh-cn/docs/guide/workbench/workflow/tool.md

@@ -58,7 +58,7 @@ weight: 236
 
 #### 用途
 
-默认清空下，工具调用节点，在决定调用工具后，会将工具运行的结果，返回给AI，让 AI 对工具运行的结果进行总结输出。有时候，如果你不需要 AI 进行进一步的总结输出，可以使用该节点，将其接入对于工具流程的末尾。
+默认情况下，工具调用节点，在决定调用工具后，会将工具运行的结果，返回给AI，让 AI 对工具运行的结果进行总结输出。有时候，如果你不需要 AI 进行进一步的总结输出，可以使用该节点，将其接入对于工具流程的末尾。
 
 如下图，在执行知识库搜索后，发送给了 HTTP 请求，搜索将不会返回搜索的结果给工具调用进行 AI 总结。
 

docSite/content/zh-cn/docs/shopping_cart/intro.md

@@ -19,17 +19,20 @@ FastGPT 商业版是基于 FastGPT 开源版的增强版本，增加了一些独
 | 应用管理与高级编排 | ✅ | ✅ | ✅ |
 | 文档知识库 | ✅ | ✅ | ✅ |
 | 外部使用 | ✅ | ✅ | ✅ |
+| API 知识库 | ✅ | ✅ | ✅ |
 | 最大应用数量 | 500 | 无限制 | 由付费套餐决定 |
 | 最大知识库数量（单个知识库内容无限制） | 30 | 无限制 | 由付费套餐决定 |
 | 自定义版权信息 | ❌ | ✅ | 设计中 |
 | 多租户与支付 | ❌ | ✅ | ✅ |
-| 团队空间 | ❌ | ✅ | ✅ |
+| 团队空间 & 权限 | ❌ | ✅ | ✅ |
 | 应用发布安全配置 | ❌ | ✅ | ✅ |
 | 内容审核 | ❌ | ✅ | ✅ |
 | web站点同步 | ❌ | ✅ | ✅ |
-| 管理后台 | ❌ | ✅ | 不需要 |
+| 主流文档库接入（目前支持：语雀、飞书） | ❌ | ✅ | ✅ |
 | 增强训练模式 | ❌ | ✅ | ✅ |
 | 第三方应用快速接入（飞书、公众号） | ❌ | ✅ | ✅ |
+| 管理后台 | ❌ | ✅ | 不需要 |
+| SSO 登录（可自定义，也可使用内置：Github、公众号、钉钉、谷歌等） | ❌ | ✅ | 不需要 |
 | 图片知识库 | ❌ | 设计中 | 设计中 |
 | 对话日志运营分析 | ❌ | 设计中 | 设计中 |
 | 完整商业授权 | ❌ | ✅ | ✅ |
@@ -50,8 +53,8 @@ FastGPT 商业版软件根据不同的部署方式，分为 3 类收费模式。
 {{< table "table-hover table-striped-columns" >}}
 | 部署方式 | 特有服务 | 上线时长 | 标品价格 |
 | ---- | ---- | ---- | ---- |
-| Sealos全托管 |  1. 有效期内免费升级。<br>2. 免运维服务&数据库。 |  半天  | 5000元起/月（3个月起）<br>或<br>50000元起/年  |
-| 自有服务器部署 |  1. 6个版本的升级服务。 | 14天内 |  具体价格可[联系咨询](https://fael3z0zfze.feishu.cn/share/base/form/shrcnRxj3utrzjywsom96Px4sud) |
+| Sealos全托管 |  1. 有效期内免费升级。<br>2. 免运维服务&数据库。 |  半天  | 6000元起/月（3个月起）<br>或<br>60000元起/年  |
+| 自有服务器部署 |  1. 6个版本免费升级支持。 | 14天内 |  具体价格可[联系咨询](https://fael3z0zfze.feishu.cn/share/base/form/shrcnRxj3utrzjywsom96Px4sud) |
 {{< /table >}}
 
 {{% alert icon="🤖 " context="success" %}}
@@ -62,6 +65,10 @@ FastGPT 商业版软件根据不同的部署方式，分为 3 类收费模式。
 - 高可用版适合对外提供在线服务，包含可视化监控、多副本、负载均衡、数据库自动备份等生产环境的基础设施。
 {{% /alert %}}
 
+## 联系方式
+
+请填写[咨询问卷](https://fael3z0zfze.feishu.cn/share/base/form/shrcnRxj3utrzjywsom96Px4sud)，我们会尽快与您联系。
+
 
 ## 技术支持
 
@@ -79,9 +86,6 @@ FastGPT 商业版软件根据不同的部署方式，分为 3 类收费模式。
 
 跨版本更新或复杂更新可参考文档自行更新；或付费支持，标准与技术服务费一致。
 
-## 联系方式
-
-请填写[咨询问卷](https://fael3z0zfze.feishu.cn/share/base/form/shrcnRxj3utrzjywsom96Px4sud)，我们会尽快与您联系。
 
 ## QA
 
@@ -95,8 +99,14 @@ FastGPT 商业版软件根据不同的部署方式，分为 3 类收费模式。
 
    可以修改开源版部分代码，不支持修改商业版镜像。完整版本=开源版+商业版镜像，所以是可以修改部分内容的。但是如果二开了，后续则需要自己进行代码合并升级。
 
-## Sealos 费用
+### Sealos 运行费用
 
 Sealos 云服务属于按量计费，下面是它的价格表：
 
-![](/imgs/sealos_price.jpg)
+![alt text](/imgs/image-58.png)
+
+## 管理后台部分截图
+
+| | | |
+| ---- | ---- | ---- |
+| ![alt text](/imgs/image-55.png) | ![alt text](/imgs/image-56.png) | ![alt text](/imgs/image-57.png) |

docSite/content/zh-cn/docs/use-cases/external-integration/official_account.md

@@ -17,9 +17,11 @@ weight: 506
 
 ![图片](/imgs/offiaccount-1.png)
 
-## 2. 登录微信公众平台，获取 AppID 、 Secret和Token
+## 2. 获取 AppID 、 Secret和Token
 
-### 1. https://mp.weixin.qq.com 登录微信公众平台，选择您的公众号。
+### 1. 登录微信公众平台，选择您的公众号。
+
+打开微信公众号官网：https://mp.weixin.qq.com
 
 **只支持通过验证的公众号，未通过验证的公众号暂不支持。**
 
@@ -28,6 +30,7 @@ weight: 506
 ![图片](/imgs/offiaccount-2.png)
 
 ### 2. 把3个参数填入 FastGPT 配置弹窗中。
+
 ![图片](/imgs/offiaccount-3.png)
 
 ## 3. 在 IP 白名单中加入 FastGPT 的 IP
@@ -36,7 +39,7 @@ weight: 506
 
 私有部署的用户可自行查阅自己的 IP 地址。
 
-海外版用户（cloud.tryfastgpt.ai)可以填写下面的 IP 白名单：
+海外版用户（cloud.tryfastgpt.ai）可以填写下面的 IP 白名单：
 
 ```
 35.240.227.100

files/docker/docker-compose-milvus.yml

@@ -114,15 +114,15 @@ services:
   # fastgpt
   sandbox:
     container_name: sandbox
-    image: ghcr.io/labring/fastgpt-sandbox:v4.8.15 # git
-    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt-sandbox:v4.8.15 # 阿里云
+    image: ghcr.io/labring/fastgpt-sandbox:v4.8.17 # git
+    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt-sandbox:v4.8.17 # 阿里云
     networks:
       - fastgpt
     restart: always
   fastgpt:
     container_name: fastgpt
-    image: ghcr.io/labring/fastgpt:v4.8.15-fix2 # git
-    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt:v4.8.15-fix2 # 阿里云
+    image: ghcr.io/labring/fastgpt:v4.8.17 # git
+    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt:v4.8.17 # 阿里云
     ports:
       - 3000:3000
     networks:

files/docker/docker-compose-pgvector.yml

@@ -72,15 +72,15 @@ services:
   # fastgpt
   sandbox:
     container_name: sandbox
-    image: ghcr.io/labring/fastgpt-sandbox:v4.8.15 # git
-    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt-sandbox:v4.8.15 # 阿里云
+    image: ghcr.io/labring/fastgpt-sandbox:v4.8.17 # git
+    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt-sandbox:v4.8.17 # 阿里云
     networks:
       - fastgpt
     restart: always
   fastgpt:
     container_name: fastgpt
-    image: ghcr.io/labring/fastgpt:v4.8.15-fix2 # git
-    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt:v4.8.15-fix2 # 阿里云
+    image: ghcr.io/labring/fastgpt:v4.8.17 # git
+    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt:v4.8.17 # 阿里云
     ports:
       - 3000:3000
     networks:

files/docker/docker-compose-zilliz.yml

@@ -53,15 +53,15 @@ services:
         wait $$!
   sandbox:
     container_name: sandbox
-    image: ghcr.io/labring/fastgpt-sandbox:v4.8.15 # git
-    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt-sandbox:v4.8.15 # 阿里云
+    image: ghcr.io/labring/fastgpt-sandbox:v4.8.17 # git
+    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt-sandbox:v4.8.17 # 阿里云
     networks:
       - fastgpt
     restart: always
   fastgpt:
     container_name: fastgpt
-    image: ghcr.io/labring/fastgpt:v4.8.15-fix2 # git
-    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt:v4.8.15-fix2 # 阿里云
+    image: ghcr.io/labring/fastgpt:v4.8.17 # git
+    # image: registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt:v4.8.17 # 阿里云
     ports:
       - 3000:3000
     networks:

packages/global/common/system/types/index.d.ts

@@ -5,7 +5,7 @@ import type {
   LLMModelItemType,
   VectorModelItemType,
   AudioSpeechModels,
-  WhisperModelType,
+  STTModelType,
   ReRankModelItemType
 } from '../../../core/ai/model.d';
 import { SubTypeEnum } from '../../../support/wallet/sub/constants';
@@ -18,6 +18,14 @@ export type NavbarItemType = {
   isActive: boolean;
 };
 
+export type ExternalProviderWorkflowVarType = {
+  name: string;
+  key: string;
+  intro: string;
+  isOpen: boolean;
+  url?: string;
+};
+
 /* fastgpt main */
 export type FastGPTConfigFileType = {
   feConfigs: FastGPTFeConfigsType;
@@ -27,7 +35,7 @@ export type FastGPTConfigFileType = {
   vectorModels: VectorModelItemType[];
   reRankModels: ReRankModelItemType[];
   audioSpeechModels: AudioSpeechModelType[];
-  whisperModel: WhisperModelType;
+  whisperModel: STTModelType;
 };
 
 export type FastGPTFeConfigsType = {
@@ -84,6 +92,7 @@ export type FastGPTFeConfigsType = {
   uploadFileMaxSize?: number;
   lafEnv?: string;
   navbarItems?: NavbarItemType[];
+  externalProviderWorkflowVariables?: ExternalProviderWorkflowVarType[];
 };
 
 export type SystemEnvType = {

packages/global/core/ai/model.d.ts

@@ -1,6 +1,13 @@
 import type { ModelProviderIdType } from './provider';
 
-export type LLMModelItemType = {
+type PriceType = {
+  charsPointsPrice?: number; // 1k chars=n points; 60s=n points;
+
+  // If inputPrice is set, the input-output charging scheme is adopted
+  inputPrice?: number; // 1k tokens=n points
+  outputPrice?: number; // 1k tokens=n points
+};
+export type LLMModelItemType = PriceType & {
   provider: ModelProviderIdType;
   model: string;
   name: string;
@@ -10,8 +17,6 @@ export type LLMModelItemType = {
   quoteMaxToken: number;
   maxTemperature: number;
 
-  charsPointsPrice: number; // 1k chars=n points
-
   censor?: boolean;
   vision?: boolean;
 
@@ -33,13 +38,12 @@ export type LLMModelItemType = {
   fieldMap?: Record<string, string>;
 };
 
-export type VectorModelItemType = {
+export type VectorModelItemType = PriceType & {
   provider: ModelProviderIdType;
   model: string; // model name
   name: string; // show name
   avatar?: string;
   defaultToken: number; // split text default token
-  charsPointsPrice: number; // 1k tokens=n points
   maxToken: number; // model max token
   weight: number; // training weight
   hidden?: boolean; // Disallow creation
@@ -48,23 +52,22 @@ export type VectorModelItemType = {
   queryConfig?: Record<string, any>; // Custom parameters for query
 };
 
-export type ReRankModelItemType = {
+export type ReRankModelItemType = PriceType & {
   model: string;
   name: string;
-  charsPointsPrice: number;
   requestUrl: string;
   requestAuth: string;
 };
 
-export type AudioSpeechModelType = {
+export type AudioSpeechModelType = PriceType & {
+  provider: ModelProviderIdType;
   model: string;
   name: string;
-  charsPointsPrice: number;
   voices: { label: string; value: string; bufferId: string }[];
 };
 
-export type WhisperModelType = {
+export type STTModelType = PriceType & {
+  provider: ModelProviderIdType;
   model: string;
   name: string;
-  charsPointsPrice: number; // 60s = n points
 };

packages/global/core/ai/model.ts

@@ -1,4 +1,5 @@
-import type { LLMModelItemType, VectorModelItemType } from './model.d';
+import { i18nT } from '../../../web/i18n/utils';
+import type { LLMModelItemType, STTModelType, VectorModelItemType } from './model.d';
 import { getModelProvider, ModelProviderIdType } from './provider';
 
 export const defaultQAModels: LLMModelItemType[] = [
@@ -35,6 +36,13 @@ export const defaultVectorModels: VectorModelItemType[] = [
   }
 ];
 
+export const defaultWhisperModel: STTModelType = {
+  provider: 'OpenAI',
+  model: 'whisper-1',
+  name: 'whisper-1',
+  charsPointsPrice: 0
+};
+
 export const getModelFromList = (
   modelList: { provider: ModelProviderIdType; name: string; model: string }[],
   model: string
@@ -46,3 +54,16 @@ export const getModelFromList = (
     avatar: provider.avatar
   };
 };
+
+export enum ModelTypeEnum {
+  chat = 'chat',
+  embedding = 'embedding',
+  tts = 'tts',
+  stt = 'stt'
+}
+export const modelTypeList = [
+  { label: i18nT('common:model.type.chat'), value: ModelTypeEnum.chat },
+  { label: i18nT('common:model.type.embedding'), value: ModelTypeEnum.embedding },
+  { label: i18nT('common:model.type.tts'), value: ModelTypeEnum.tts },
+  { label: i18nT('common:model.type.stt'), value: ModelTypeEnum.stt }
+];

packages/global/core/ai/provider.ts

@@ -4,19 +4,25 @@ export type ModelProviderIdType =
   | 'OpenAI'
   | 'Claude'
   | 'Gemini'
+  | 'Meta'
   | 'MistralAI'
+  | 'Groq'
+  | 'AliCloud'
   | 'Qwen'
   | 'Doubao'
   | 'ChatGLM'
   | 'DeepSeek'
+  | 'Ernie'
   | 'Moonshot'
   | 'MiniMax'
   | 'SparkDesk'
   | 'Hunyuan'
   | 'Baichuan'
+  | 'StepFun'
   | 'Yi'
-  | 'Ernie'
   | 'Ollama'
+  | 'BAAI'
+  | 'FishAudio'
   | 'Other';
 
 export type ModelProviderType = {
@@ -41,10 +47,25 @@ export const ModelProviderList: ModelProviderType[] = [
     name: 'Gemini',
     avatar: 'model/gemini'
   },
+  {
+    id: 'Meta',
+    name: 'Meta',
+    avatar: 'model/meta'
+  },
   {
     id: 'MistralAI',
     name: 'MistralAI',
-    avatar: 'model/huggingface'
+    avatar: 'model/mistral'
+  },
+  {
+    id: 'Groq',
+    name: 'Groq',
+    avatar: 'model/groq'
+  },
+  {
+    id: 'AliCloud',
+    name: i18nT('common:model_alicloud'),
+    avatar: 'model/alicloud'
   },
   {
     id: 'Qwen',
@@ -61,6 +82,11 @@ export const ModelProviderList: ModelProviderType[] = [
     name: i18nT('common:model_chatglm'),
     avatar: 'model/chatglm'
   },
+  {
+    id: 'Ernie',
+    name: i18nT('common:model_ernie'),
+    avatar: 'model/ernie'
+  },
   {
     id: 'DeepSeek',
     name: 'DeepSeek',
@@ -91,21 +117,32 @@ export const ModelProviderList: ModelProviderType[] = [
     name: i18nT('common:model_baichuan'),
     avatar: 'model/baichuan'
   },
+  {
+    id: 'StepFun',
+    name: i18nT('common:model_stepfun'),
+    avatar: 'model/stepfun'
+  },
   {
     id: 'Yi',
     name: i18nT('common:model_yi'),
     avatar: 'model/yi'
   },
-  {
-    id: 'Ernie',
-    name: i18nT('common:model_ernie'),
-    avatar: 'model/ernie'
-  },
+
   {
     id: 'Ollama',
     name: 'Ollama',
     avatar: 'model/ollama'
   },
+  {
+    id: 'BAAI',
+    name: i18nT('common:model_baai'),
+    avatar: 'model/BAAI'
+  },
+  {
+    id: 'FishAudio',
+    name: 'FishAudio',
+    avatar: 'model/fishaudio'
+  },
   {
     id: 'Other',
     name: i18nT('common:model_other'),
@@ -113,7 +150,7 @@ export const ModelProviderList: ModelProviderType[] = [
   }
 ];
 export const ModelProviderMap = Object.fromEntries(
-  ModelProviderList.map((item) => [item.id, item])
+  ModelProviderList.map((item, index) => [item.id, { ...item, order: index }])
 );
 
 export const getModelProvider = (provider: ModelProviderIdType) => {

packages/global/core/app/type.d.ts

@@ -13,6 +13,7 @@ import { StoreEdgeItemType } from '../workflow/type/edge';
 import { AppPermission } from '../../support/permission/app/controller';
 import { ParentIdType } from '../../common/parentFolder/type';
 import { FlowNodeInputTypeEnum } from 'core/workflow/node/constant';
+import { WorkflowTemplateBasicType } from '@fastgpt/global/core/workflow/type';
 
 export type AppSchema = {
   _id: string;
@@ -73,8 +74,8 @@ export type AppSimpleEditFormType = {
   aiSettings: {
     model: string;
     systemPrompt?: string | undefined;
-    temperature: number;
-    maxToken: number;
+    temperature?: number;
+    maxToken?: number;
     isResponseAnswerText: boolean;
     maxHistories: number;
   };
@@ -109,8 +110,8 @@ export type AppChatConfigType = {
 };
 export type SettingAIDataType = {
   model: string;
-  temperature: number;
-  maxToken: number;
+  temperature?: number;
+  maxToken?: number;
   isResponseAnswerText?: boolean;
   maxHistories?: number;
   [NodeInputKeyEnum.aiChatVision]?: boolean; // Is open vision mode
@@ -184,3 +185,28 @@ export type SystemPluginListItemType = {
   name: string;
   avatar: string;
 };
+
+export type AppTemplateSchemaType = {
+  templateId: string;
+  name: string;
+  intro: string;
+  avatar: string;
+  tags: string[];
+  type: string;
+  author?: string;
+  isActive?: boolean;
+  userGuide?: {
+    type: 'markdown' | 'link';
+    content?: string;
+    link?: string;
+  };
+  isQuickTemplate?: boolean;
+  order?: number;
+  workflow: WorkflowTemplateBasicType;
+};
+
+export type TemplateTypeSchemaType = {
+  typeName: string;
+  typeId: string;
+  typeOrder: number;
+};

packages/global/core/dataset/collection/utils.ts

@@ -1,9 +1,7 @@
 import { DatasetCollectionTypeEnum, TrainingModeEnum, TrainingTypeMap } from '../constants';
-import { CollectionWithDatasetType, DatasetCollectionSchemaType } from '../type';
+import { DatasetCollectionSchemaType } from '../type';
 
-export const getCollectionSourceData = (
-  collection?: CollectionWithDatasetType | DatasetCollectionSchemaType
-) => {
+export const getCollectionSourceData = (collection?: DatasetCollectionSchemaType) => {
   return {
     sourceId:
       collection?.fileId ||

packages/global/core/dataset/type.d.ts

@@ -133,11 +133,8 @@ export type DatasetTrainingSchemaType = {
   indexes: Omit<DatasetDataIndexItemType, 'dataId'>[];
 };
 
-export type CollectionWithDatasetType = Omit<DatasetCollectionSchemaType, 'datasetId'> & {
-  datasetId: DatasetSchemaType;
-};
-export type DatasetDataWithCollectionType = Omit<DatasetDataSchemaType, 'collectionId'> & {
-  collectionId: DatasetCollectionSchemaType;
+export type CollectionWithDatasetType = DatasetCollectionSchemaType & {
+  dataset: DatasetSchemaType;
 };
 
 /* ================= dataset ===================== */

packages/global/core/workflow/runtime/type.d.ts

@@ -21,13 +21,20 @@ import { ReadFileNodeResponse } from '../template/system/readFiles/type';
 import { UserSelectOptionType } from '../template/system/userSelect/type';
 import { WorkflowResponseType } from '../../../../service/core/workflow/dispatch/type';
 import { AiChatQuoteRoleType } from '../template/system/aiChat/type';
+import { LafAccountType, OpenaiAccountType } from '../../../support/user/team/type';
+
+export type ExternalProviderType = {
+  openaiAccount?: OpenaiAccountType;
+  externalWorkflowVariables?: Record<string, string>;
+};
 
 /* workflow props */
 export type ChatDispatchProps = {
   res?: NextApiResponse;
   requestOrigin?: string;
   mode: 'test' | 'chat' | 'debug';
-  user: UserModelSchema;
+  timezone: string;
+  externalProvider: ExternalProviderType;
 
   runningAppInfo: {
     id: string; // May be the id of the system plug-in (cannot be used directly to look up the table)
@@ -100,7 +107,9 @@ export type DispatchNodeResponseType = {
   mergeSignId?: string;
 
   // bill
-  tokens?: number;
+  tokens?: number; // deprecated
+  inputTokens?: number;
+  outputTokens?: number;
   model?: string;
   contextTotalLen?: number;
   totalPoints?: number;
@@ -150,6 +159,8 @@ export type DispatchNodeResponseType = {
 
   // tool
   toolCallTokens?: number;
+  toolCallInputTokens?: number;
+  toolCallOutputTokens?: number;
   toolDetail?: ChatHistoryItemResType[];
   toolStop?: boolean;
 
@@ -201,13 +212,14 @@ export type DispatchNodeResultType<T = {}> = {
 export type AIChatNodeProps = {
   [NodeInputKeyEnum.aiModel]: string;
   [NodeInputKeyEnum.aiSystemPrompt]?: string;
-  [NodeInputKeyEnum.aiChatTemperature]: number;
-  [NodeInputKeyEnum.aiChatMaxToken]: number;
+  [NodeInputKeyEnum.aiChatTemperature]?: number;
+  [NodeInputKeyEnum.aiChatMaxToken]?: number;
   [NodeInputKeyEnum.aiChatIsResponseText]: boolean;
+  [NodeInputKeyEnum.aiChatVision]?: boolean;
+
   [NodeInputKeyEnum.aiChatQuoteRole]?: AiChatQuoteRoleType;
   [NodeInputKeyEnum.aiChatQuoteTemplate]?: string;
   [NodeInputKeyEnum.aiChatQuotePrompt]?: string;
-  [NodeInputKeyEnum.aiChatVision]?: boolean;
 
   [NodeInputKeyEnum.stringQuoteText]?: string;
   [NodeInputKeyEnum.fileUrlList]?: string[];

packages/global/core/workflow/runtime/utils.ts

@@ -9,6 +9,7 @@ import { isValidReferenceValueFormat } from '../utils';
 import { FlowNodeOutputItemType, ReferenceValueType } from '../type/io';
 import { ChatItemType, NodeOutputItemType } from '../../../core/chat/type';
 import { ChatItemValueTypeEnum, ChatRoleEnum } from '../../../core/chat/constants';
+import { replaceVariable } from '../../../common/string/tools';
 
 export const getMaxHistoryLimitFromNodes = (nodes: StoreNodeItemType[]): number => {
   let limit = 10;
@@ -317,6 +318,8 @@ export function replaceEditorVariable({
 }) {
   if (typeof text !== 'string') return text;
 
+  text = replaceVariable(text, variables);
+
   const variablePattern = /\{\{\$([^.]+)\.([^$]+)\$\}\}/g;
   const matches = [...text.matchAll(variablePattern)];
   if (matches.length === 0) return text;

packages/global/core/workflow/template/system/workflowStart.ts

@@ -30,7 +30,6 @@ export const WorkflowStart: FlowNodeTemplateType = {
   intro: '',
   forbidDelete: true,
   unique: true,
-  courseUrl: '/docs/guide/workbench/workflow/input/',
   version: '481',
   inputs: [{ ...Input_Template_UserChatInput, toolDescription: i18nT('workflow:user_question') }],
   outputs: [

packages/global/core/workflow/type/node.d.ts

@@ -69,6 +69,7 @@ export type FlowNodeTemplateType = FlowNodeCommonType & {
 
   diagram?: string; // diagram url
   courseUrl?: string; // course url
+  userGuide?: string; // user guide
 };
 
 export type NodeTemplateListItemType = {
@@ -87,6 +88,7 @@ export type NodeTemplateListItemType = {
   currentCost?: number; // 当前积分消耗
   hasTokenFee?: boolean; // 是否配置积分
   instructions?: string; // 使用说明
+  courseUrl?: string; // 教程链接
 };
 
 export type NodeTemplateListType = {

packages/global/support/outLink/type.d.ts

@@ -83,11 +83,6 @@ export type OutLinkSchema<T extends OutlinkAppType = undefined> = {
   app: T;
 };
 
-// to handle MongoDB querying
-export type OutLinkWithAppType = Omit<OutLinkSchema, 'appId'> & {
-  appId: AppSchema;
-};
-
 // Edit the Outlink
 export type OutLinkEditType<T = undefined> = {
   _id?: string;

packages/global/support/permission/type.d.ts

@@ -1,5 +1,6 @@
+import { UserModelSchema } from '../user/type';
 import { RequireOnlyOne } from '../../common/type/utils';
-import { TeamMemberWithUserSchema } from '../user/team/type';
+import { TeamMemberSchema } from '../user/team/type';
 import { AuthUserTypeEnum, PermissionKeyEnum, PerResourceTypeEnum } from './constant';
 import { MemberGroupSchemaType } from './memberGroup/type';
 
@@ -31,11 +32,7 @@ export type ResourcePermissionType = {
 }>;
 
 export type ResourcePerWithTmbWithUser = Omit<ResourcePermissionType, 'tmbId'> & {
-  tmbId: TeamMemberWithUserSchema;
-};
-
-export type ResourcePerWithGroup = Omit<ResourcePermissionType, 'groupId'> & {
-  groupId: MemberGroupSchemaType;
+  tmbId: TeamMemberSchema & { user: UserModelSchema };
 };
 
 export type PermissionSchemaType = {

packages/global/support/user/team/controller.d.ts

@@ -1,6 +1,6 @@
 import { PermissionValueType } from '../../permission/type';
 import { TeamMemberRoleEnum } from './constant';
-import { LafAccountType, TeamMemberSchema } from './type';
+import { LafAccountType, TeamMemberSchema, ThirdPartyAccountType } from './type';
 
 export type AuthTeamRoleProps = {
   teamId: string;
@@ -11,14 +11,13 @@ export type CreateTeamProps = {
   name: string;
   avatar?: string;
   defaultTeam?: boolean;
-  lafAccount?: LafAccountType;
   memberName?: string;
 };
-export type UpdateTeamProps = {
+export type UpdateTeamProps = Omit<ThirdPartyAccountType, 'externalWorkflowVariable'> & {
   name?: string;
   avatar?: string;
   teamDomain?: string;
-  lafAccount?: null | LafAccountType;
+  externalWorkflowVariable?: { key: string; value: string };
 };
 
 /* ------------- member ----------- */

packages/global/support/user/team/type.d.ts

@@ -4,6 +4,12 @@ import { LafAccountType } from './type';
 import { PermissionValueType, ResourcePermissionType } from '../../permission/type';
 import { TeamPermission } from '../../permission/user/controller';
 
+export type ThirdPartyAccountType = {
+  lafAccount?: LafAccountType;
+  openaiAccount?: OpenaiAccountType;
+  externalWorkflowVariables?: Record<string, string>;
+};
+
 export type TeamSchema = {
   _id: string;
   name: string;
@@ -16,9 +22,8 @@ export type TeamSchema = {
     lastExportDatasetTime: Date;
     lastWebsiteSyncTime: Date;
   };
-  lafAccount: LafAccountType;
   notificationAccount?: string;
-};
+} & ThirdPartyAccountType;
 
 export type tagsType = {
   label: string;
@@ -42,16 +47,9 @@ export type TeamMemberSchema = {
   defaultTeam: boolean;
 };
 
-export type TeamMemberWithUserSchema = Omit<TeamMemberSchema, 'userId'> & {
-  userId: UserModelSchema;
-};
-
-export type TeamMemberWithTeamSchema = Omit<TeamMemberSchema, 'teamId'> & {
-  teamId: TeamSchema;
-};
-
-export type TeamMemberWithTeamAndUserSchema = Omit<TeamMemberWithTeamSchema, 'userId'> & {
-  userId: UserModelSchema;
+export type TeamMemberWithTeamAndUserSchema = TeamMemberSchema & {
+  team: TeamSchema;
+  user: UserModelSchema;
 };
 
 export type TeamTmbItemType = {
@@ -66,10 +64,9 @@ export type TeamTmbItemType = {
   defaultTeam: boolean;
   role: `${TeamMemberRoleEnum}`;
   status: `${TeamMemberStatusEnum}`;
-  lafAccount?: LafAccountType;
   notificationAccount?: string;
   permission: TeamPermission;
-};
+} & ThirdPartyAccountType;
 
 export type TeamMemberItemType = {
   userId: string;
@@ -88,11 +85,16 @@ export type TeamTagItemType = {
 };
 
 export type LafAccountType = {
-  token: string;
   appid: string;
+  token: string;
   pat: string;
 };
 
+export type OpenaiAccountType = {
+  key: string;
+  baseUrl: string;
+};
+
 export type TeamInvoiceHeaderType = {
   teamName: string;
   unifiedCreditCode: string;

packages/global/support/user/type.d.ts

@@ -14,10 +14,6 @@ export type UserModelSchema = {
   timezone: string;
   status: `${UserStatusEnum}`;
   lastLoginTmbId?: string;
-  openaiAccount?: {
-    key: string;
-    baseUrl: string;
-  };
   fastgpt_sem?: {
     keyword: string;
   };
@@ -29,7 +25,6 @@ export type UserType = {
   avatar: string;
   timezone: string;
   promotionRate: UserModelSchema['promotionRate'];
-  openaiAccount: UserModelSchema['openaiAccount'];
   team: TeamTmbItemType;
   standardInfo?: standardInfoType;
   notificationAccount?: string;

packages/global/support/wallet/bill/type.d.ts

@@ -23,7 +23,8 @@ export type BillSchemaType = {
 };
 
 export type ChatNodeUsageType = {
-  tokens?: number;
+  inputTokens?: number;
+  outputTokens?: number;
   totalPoints: number;
   moduleName: string;
   model?: string;

packages/global/support/wallet/usage/type.d.ts

@@ -2,9 +2,13 @@ import { CreateUsageProps } from './api';
 import { UsageSourceEnum } from './constants';
 
 export type UsageListItemCountType = {
-  tokens?: number;
+  inputTokens?: number;
+  outputTokens?: number;
   charsLength?: number;
   duration?: number;
+
+  // deprecated
+  tokens?: number;
 };
 export type UsageListItemType = UsageListItemCountType & {
   moduleName: string;

packages/plugins/src/feishu/template.json

@@ -2,7 +2,7 @@
   "author": "",
   "version": "488",
   "name": "飞书 webhook",
-  "avatar": "/appMarketTemplates/plugin-feishu/avatar.svg",
+  "avatar": "core/app/templates/plugin-feishu",
   "intro": "向飞书机器人发起 webhook 请求。",
   "courseUrl": "https://open.feishu.cn/document/client-docs/bot-v3/add-custom-bot#f62e72d5",
   "showStatus": false,

packages/plugins/src/searchXNG/template.json

@@ -55,8 +55,7 @@
             "maxFiles": 5,
             "canSelectFile": true,
             "canSelectImg": true,
-            "required": true,
-            "toolDescription": "部署的searXNG服务的链接"
+            "required": true
           }
         ],
         "outputs": [

packages/service/common/system/config/controller.ts

@@ -4,7 +4,12 @@ import { FastGPTConfigFileType } from '@fastgpt/global/common/system/types';
 import { FastGPTProUrl } from '../constants';
 
 export const getFastGPTConfigFromDB = async () => {
-  if (!FastGPTProUrl) return {} as FastGPTConfigFileType;
+  if (!FastGPTProUrl) {
+    return {
+      config: {} as FastGPTConfigFileType,
+      configId: undefined
+    };
+  }
 
   const res = await MongoSystemConfigs.findOne({
     type: SystemConfigsTypeEnum.fastgpt
@@ -14,5 +19,8 @@ export const getFastGPTConfigFromDB = async () => {
 
   const config = res?.value || {};
 
-  return config as FastGPTConfigFileType;
+  return {
+    configId: res ? String(res._id) : undefined,
+    config: config as FastGPTConfigFileType
+  };
 };

packages/service/common/system/tools.ts

@@ -15,6 +15,9 @@ export const initFastGPTConfig = (config?: FastGPTConfigFileType) => {
   global.subPlans = config.subPlans;
 
   global.llmModels = config.llmModels;
+  global.llmModelPriceType = global.llmModels.some((item) => typeof item.inputPrice === 'number')
+    ? 'IO'
+    : 'Tokens';
   global.vectorModels = config.vectorModels;
   global.audioSpeechModels = config.audioSpeechModels;
   global.whisperModel = config.whisperModel;

packages/service/core/ai/config.ts

@@ -1,4 +1,3 @@
-import type { UserModelSchema } from '@fastgpt/global/support/user/type';
 import OpenAI from '@fastgpt/global/core/ai';
 import {
   ChatCompletionCreateParamsNonStreaming,
@@ -7,13 +6,11 @@ import {
 import { getErrText } from '@fastgpt/global/common/error/utils';
 import { addLog } from '../../common/system/log';
 import { i18nT } from '../../../web/i18n/utils';
+import { OpenaiAccountType } from '@fastgpt/global/support/user/team/type';
 
 export const openaiBaseUrl = process.env.OPENAI_BASE_URL || 'https://api.openai.com/v1';
 
-export const getAIApi = (props?: {
-  userKey?: UserModelSchema['openaiAccount'];
-  timeout?: number;
-}) => {
+export const getAIApi = (props?: { userKey?: OpenaiAccountType; timeout?: number }) => {
   const { userKey, timeout } = props || {};
 
   const baseUrl =
@@ -29,7 +26,7 @@ export const getAIApi = (props?: {
   });
 };
 
-export const getAxiosConfig = (props?: { userKey?: UserModelSchema['openaiAccount'] }) => {
+export const getAxiosConfig = (props?: { userKey?: OpenaiAccountType }) => {
   const { userKey } = props || {};
 
   const baseUrl =
@@ -57,7 +54,7 @@ export const createChatCompletion = async <T extends CompletionsBodyType>({
   options
 }: {
   body: T;
-  userKey?: UserModelSchema['openaiAccount'];
+  userKey?: OpenaiAccountType;
   timeout?: number;
   options?: OpenAI.RequestOptions;
 }): Promise<{

packages/service/core/ai/functions/createQuestionGuide.ts

@@ -1,6 +1,6 @@
 import type { ChatCompletionMessageParam } from '@fastgpt/global/core/ai/type.d';
 import { createChatCompletion } from '../config';
-import { countGptMessagesTokens } from '../../../common/string/tiktoken/index';
+import { countGptMessagesTokens, countPromptTokens } from '../../../common/string/tiktoken/index';
 import { loadRequestMessages } from '../../chat/utils';
 import { llmCompletionsBodyFormat } from '../utils';
 import {
@@ -20,7 +20,8 @@ export async function createQuestionGuide({
   customPrompt?: string;
 }): Promise<{
   result: string[];
-  tokens: number;
+  inputTokens: number;
+  outputTokens: number;
 }> {
   const concatMessages: ChatCompletionMessageParam[] = [
     ...messages,
@@ -29,6 +30,10 @@ export async function createQuestionGuide({
       content: `${customPrompt || PROMPT_QUESTION_GUIDE}\n${PROMPT_QUESTION_GUIDE_FOOTER}`
     }
   ];
+  const requestMessages = await loadRequestMessages({
+    messages: concatMessages,
+    useVision: false
+  });
 
   const { response: data } = await createChatCompletion({
     body: llmCompletionsBodyFormat(
@@ -36,10 +41,7 @@ export async function createQuestionGuide({
         model,
         temperature: 0.1,
         max_tokens: 200,
-        messages: await loadRequestMessages({
-          messages: concatMessages,
-          useVision: false
-        }),
+        messages: requestMessages,
         stream: false
       },
       model
@@ -51,13 +53,15 @@ export async function createQuestionGuide({
   const start = answer.indexOf('[');
   const end = answer.lastIndexOf(']');
 
-  const tokens = await countGptMessagesTokens(concatMessages);
+  const inputTokens = await countGptMessagesTokens(requestMessages);
+  const outputTokens = await countPromptTokens(answer);
 
   if (start === -1 || end === -1) {
     addLog.warn('Create question guide error', { answer });
     return {
       result: [],
-      tokens: 0
+      inputTokens: 0,
+      outputTokens: 0
     };
   }
 
@@ -69,14 +73,16 @@ export async function createQuestionGuide({
   try {
     return {
       result: json5.parse(jsonStr),
-      tokens
+      inputTokens,
+      outputTokens
     };
   } catch (error) {
     console.log(error);
 
     return {
       result: [],
-      tokens: 0
+      inputTokens: 0,
+      outputTokens: 0
     };
   }
 }

packages/service/core/ai/functions/queryExtension.ts

@@ -1,7 +1,7 @@
 import { replaceVariable } from '@fastgpt/global/common/string/tools';
 import { createChatCompletion } from '../config';
 import { ChatItemType } from '@fastgpt/global/core/chat/type';
-import { countGptMessagesTokens } from '../../../common/string/tiktoken/index';
+import { countGptMessagesTokens, countPromptTokens } from '../../../common/string/tiktoken/index';
 import { chatValue2RuntimePrompt } from '@fastgpt/global/core/chat/adapt';
 import { getLLMModel } from '../model';
 import { llmCompletionsBodyFormat } from '../utils';
@@ -121,7 +121,8 @@ export const queryExtension = async ({
   rawQuery: string;
   extensionQueries: string[];
   model: string;
-  tokens: number;
+  inputTokens: number;
+  outputTokens: number;
 }> => {
   const systemFewShot = chatBg
     ? `Q: 对话背景。
@@ -166,7 +167,8 @@ A: ${chatBg}
       rawQuery: query,
       extensionQueries: [],
       model,
-      tokens: 0
+      inputTokens: 0,
+      outputTokens: 0
     };
   }
 
@@ -181,7 +183,8 @@ A: ${chatBg}
       rawQuery: query,
       extensionQueries: Array.isArray(queries) ? queries : [],
       model,
-      tokens: await countGptMessagesTokens(messages)
+      inputTokens: await countGptMessagesTokens(messages),
+      outputTokens: await countPromptTokens(answer)
     };
   } catch (error) {
     addLog.error(`Query extension error`, error);
@@ -189,7 +192,8 @@ A: ${chatBg}
       rawQuery: query,
       extensionQueries: [],
       model,
-      tokens: 0
+      inputTokens: 0,
+      outputTokens: 0
     };
   }
 };

packages/service/core/ai/model.ts

@@ -4,6 +4,7 @@ export const getLLMModel = (model?: string) => {
     global.llmModels[0]
   );
 };
+
 export const getDatasetModel = (model?: string) => {
   return (
     global.llmModels

packages/service/core/ai/utils.ts

@@ -12,10 +12,12 @@ export const computedMaxToken = async ({
   model,
   filterMessages = []
 }: {
-  maxToken: number;
+  maxToken?: number;
   model: LLMModelItemType;
   filterMessages: ChatCompletionMessageParam[];
 }) => {
+  if (maxToken === undefined) return;
+
   maxToken = Math.min(maxToken, model.maxResponse);
   const tokensLimit = model.maxContext;
 
@@ -37,8 +39,6 @@ export const computedTemperature = ({
   model: LLMModelItemType;
   temperature: number;
 }) => {
-  if (temperature < 1) return temperature;
-
   temperature = +(model.maxTemperature * (temperature / 10)).toFixed(2);
   temperature = Math.max(temperature, 0.01);
 
@@ -63,12 +63,13 @@ export const llmCompletionsBodyFormat = <T extends CompletionsBodyType>(
 
   const requestBody: T = {
     ...body,
-    temperature: body.temperature
-      ? computedTemperature({
-          model: modelData,
-          temperature: body.temperature
-        })
-      : undefined,
+    temperature:
+      typeof body.temperature === 'number'
+        ? computedTemperature({
+            model: modelData,
+            temperature: body.temperature
+          })
+        : undefined,
     ...modelData?.defaultConfig
   };
 

packages/service/core/app/plugin/controller.ts

@@ -131,6 +131,7 @@ export async function getChildAppPreviewNode({
     name: app.name,
     intro: app.intro,
     courseUrl: app.courseUrl,
+    userGuide: app.userGuide,
     showStatus: app.showStatus,
     isTool: true,
     version: app.version,

packages/service/core/app/templates/templateSchema.ts

@@ -0,0 +1,51 @@
+import { AppTemplateSchemaType } from '@fastgpt/global/core/app/type';
+import { connectionMongo, getMongoModel } from '../../../common/mongo/index';
+const { Schema } = connectionMongo;
+
+export const collectionName = 'app_templates';
+
+const AppTemplateSchema = new Schema({
+  templateId: {
+    type: String,
+    required: true
+  },
+  name: {
+    type: String
+  },
+  intro: {
+    type: String
+  },
+  avatar: {
+    type: String
+  },
+  tags: {
+    type: [String],
+    default: undefined
+  },
+  type: {
+    type: String
+  },
+  isActive: {
+    type: Boolean
+  },
+  userGuide: {
+    type: Object
+  },
+  isQuickTemplate: {
+    type: Boolean
+  },
+  order: {
+    type: Number,
+    default: -1
+  },
+  workflow: {
+    type: Object
+  }
+});
+
+AppTemplateSchema.index({ templateId: 1 });
+
+export const MongoAppTemplate = getMongoModel<AppTemplateSchemaType>(
+  collectionName,
+  AppTemplateSchema
+);

packages/service/core/app/templates/templateTypeSchema.ts

@@ -0,0 +1,25 @@
+import { TemplateTypeSchemaType } from '@fastgpt/global/core/app/type';
+import { connectionMongo, getMongoModel } from '../../../common/mongo/index';
+const { Schema } = connectionMongo;
+
+export const collectionName = 'app_template_types';
+
+const TemplateTypeSchema = new Schema({
+  typeName: {
+    type: String,
+    required: true
+  },
+  typeId: {
+    type: String,
+    required: true
+  },
+  typeOrder: {
+    type: Number,
+    default: 0
+  }
+});
+
+export const MongoTemplateTypes = getMongoModel<TemplateTypeSchemaType>(
+  collectionName,
+  TemplateTypeSchema
+);

packages/service/core/app/version/schema.ts

@@ -2,10 +2,16 @@ import { connectionMongo, getMongoModel, type Model } from '../../../common/mong
 const { Schema, model, models } = connectionMongo;
 import { AppVersionSchemaType } from '@fastgpt/global/core/app/version';
 import { chatConfigType } from '../schema';
+import { TeamMemberCollectionName } from '@fastgpt/global/support/user/team/constant';
 
 export const AppVersionCollectionName = 'app_versions';
 
 const AppVersionSchema = new Schema({
+  tmbId: {
+    type: String,
+    ref: TeamMemberCollectionName,
+    required: true
+  },
   appId: {
     type: Schema.Types.ObjectId,
     ref: AppVersionCollectionName,
@@ -26,16 +32,8 @@ const AppVersionSchema = new Schema({
   chatConfig: {
     type: chatConfigType
   },
-  isPublish: {
-    type: Boolean
-  },
-  versionName: {
-    type: String,
-    default: ''
-  },
-  tmbId: {
-    type: String
-  }
+  isPublish: Boolean,
+  versionName: String
 });
 
 try {

packages/service/core/chat/utils.ts

@@ -104,9 +104,6 @@ export const loadRequestMessages = async ({
 }) => {
   // Load image to base64
   const loadImageToBase64 = async (messages: ChatCompletionContentPart[]) => {
-    if (process.env.MULTIPLE_DATA_TO_BASE64 === 'false') {
-      return messages;
-    }
     return Promise.all(
       messages.map(async (item) => {
         if (item.type === 'image_url') {
@@ -125,7 +122,7 @@ export const loadRequestMessages = async ({
 
           try {
             // If imgUrl is a local path, load image from local, and set url to base64
-            if (imgUrl.startsWith('/')) {
+            if (imgUrl.startsWith('/') || process.env.MULTIPLE_DATA_TO_BASE64 === 'true') {
               addLog.debug('Load image from local server', {
                 baseUrl: serverRequestBaseUrl,
                 requestUrl: imgUrl

packages/service/core/dataset/collection/controller.ts

@@ -4,11 +4,7 @@ import {
 } from '@fastgpt/global/core/dataset/constants';
 import type { CreateDatasetCollectionParams } from '@fastgpt/global/core/dataset/api.d';
 import { MongoDatasetCollection } from './schema';
-import {
-  CollectionWithDatasetType,
-  DatasetCollectionSchemaType,
-  DatasetSchemaType
-} from '@fastgpt/global/core/dataset/type';
+import { DatasetCollectionSchemaType, DatasetSchemaType } from '@fastgpt/global/core/dataset/type';
 import { MongoDatasetTraining } from '../training/schema';
 import { MongoDatasetData } from '../data/schema';
 import { delImgByRelatedId } from '../../../common/file/image/controller';
@@ -230,7 +226,7 @@ export const delCollectionRelatedSource = async ({
   collections,
   session
 }: {
-  collections: (CollectionWithDatasetType | DatasetCollectionSchemaType)[];
+  collections: DatasetCollectionSchemaType[];
   session: ClientSession;
 }) => {
   if (collections.length === 0) return;
@@ -264,7 +260,7 @@ export async function delCollection({
   session,
   delRelatedSource
 }: {
-  collections: (CollectionWithDatasetType | DatasetCollectionSchemaType)[];
+  collections: DatasetCollectionSchemaType[];
   session: ClientSession;
   delRelatedSource: boolean;
 }) {
@@ -274,16 +270,7 @@ export async function delCollection({
 
   if (!teamId) return Promise.reject('teamId is not exist');
 
-  const datasetIds = Array.from(
-    new Set(
-      collections.map((item) => {
-        if (typeof item.datasetId === 'string') {
-          return String(item.datasetId);
-        }
-        return String(item.datasetId._id);
-      })
-    )
-  );
+  const datasetIds = Array.from(new Set(collections.map((item) => String(item.datasetId))));
   const collectionIds = collections.map((item) => String(item._id));
 
   // delete training data
@@ -324,7 +311,7 @@ export async function delOnlyCollection({
   collections,
   session
 }: {
-  collections: (CollectionWithDatasetType | DatasetCollectionSchemaType)[];
+  collections: DatasetCollectionSchemaType[];
   session: ClientSession;
 }) {
   if (collections.length === 0) return;
@@ -333,16 +320,7 @@ export async function delOnlyCollection({
 
   if (!teamId) return Promise.reject('teamId is not exist');
 
-  const datasetIds = Array.from(
-    new Set(
-      collections.map((item) => {
-        if (typeof item.datasetId === 'string') {
-          return String(item.datasetId);
-        }
-        return String(item.datasetId._id);
-      })
-    )
-  );
+  const datasetIds = Array.from(new Set(collections.map((item) => String(item.datasetId))));
   const collectionIds = collections.map((item) => String(item._id));
 
   // delete training data

packages/service/core/dataset/collection/schema.ts

@@ -100,6 +100,13 @@ const DatasetCollectionSchema = new Schema({
   }
 });
 
+DatasetCollectionSchema.virtual('dataset', {
+  ref: DatasetCollectionName,
+  localField: 'datasetId',
+  foreignField: '_id',
+  justOne: true
+});
+
 try {
   // auth file
   DatasetCollectionSchema.index({ teamId: 1, fileId: 1 });

packages/service/core/dataset/collection/utils.ts

@@ -130,7 +130,7 @@ export const collectionTagsToTagLabel = async ({
 };
 
 export const syncCollection = async (collection: CollectionWithDatasetType) => {
-  const dataset = collection.datasetId;
+  const dataset = collection.dataset;
 
   if (
     collection.type !== DatasetCollectionTypeEnum.link &&
@@ -183,7 +183,7 @@ export const syncCollection = async (collection: CollectionWithDatasetType) => {
         teamId: collection.teamId,
         tmbId: collection.tmbId,
         name: collection.name,
-        datasetId: collection.datasetId._id,
+        datasetId: collection.datasetId,
         parentId: collection.parentId,
         type: collection.type,
 

packages/service/core/dataset/controller.ts

@@ -1,4 +1,4 @@
-import { CollectionWithDatasetType, DatasetSchemaType } from '@fastgpt/global/core/dataset/type';
+import { DatasetSchemaType } from '@fastgpt/global/core/dataset/type';
 import { MongoDatasetCollection } from './collection/schema';
 import { MongoDataset } from './schema';
 import { delCollectionRelatedSource } from './collection/controller';
@@ -49,9 +49,9 @@ export async function findDatasetAndAllChildren({
 }
 
 export async function getCollectionWithDataset(collectionId: string) {
-  const data = (await MongoDatasetCollection.findById(collectionId)
-    .populate('datasetId')
-    .lean()) as CollectionWithDatasetType;
+  const data = await MongoDatasetCollection.findById(collectionId)
+    .populate<{ dataset: DatasetSchemaType }>('dataset')
+    .lean();
   if (!data) {
     return Promise.reject('Collection is not exist');
   }
@@ -71,19 +71,10 @@ export async function delDatasetRelevantData({
   const teamId = datasets[0].teamId;
 
   if (!teamId) {
-    return Promise.reject('teamId is required');
+    return Promise.reject('TeamId is required');
   }
 
-  const datasetIds = datasets.map((item) => String(item._id));
-
-  // Get _id, teamId, fileId, metadata.relatedImgId for all collections
-  const collections = await MongoDatasetCollection.find(
-    {
-      teamId,
-      datasetId: { $in: datasetIds }
-    },
-    '_id teamId datasetId fileId metadata'
-  ).lean();
+  const datasetIds = datasets.map((item) => item._id);
 
   // delete training data
   await MongoDatasetTraining.deleteMany({
@@ -91,20 +82,27 @@ export async function delDatasetRelevantData({
     datasetId: { $in: datasetIds }
   });
 
-  // image and file
-  await delCollectionRelatedSource({ collections, session });
-
-  // delete dataset.datas
-  await MongoDatasetData.deleteMany({ teamId, datasetId: { $in: datasetIds } }, { session });
-
-  // delete collections
-  await MongoDatasetCollection.deleteMany(
+  // Get _id, teamId, fileId, metadata.relatedImgId for all collections
+  const collections = await MongoDatasetCollection.find(
     {
       teamId,
       datasetId: { $in: datasetIds }
     },
+    '_id teamId datasetId fileId metadata',
     { session }
-  );
+  ).lean();
+
+  // image and file
+  await delCollectionRelatedSource({ collections, session });
+
+  // delete collections
+  await MongoDatasetCollection.deleteMany({
+    teamId,
+    datasetId: { $in: datasetIds }
+  }).session(session);
+
+  // delete dataset.datas(Not need session)
+  await MongoDatasetData.deleteMany({ teamId, datasetId: { $in: datasetIds } });
 
   // no session delete: delete files, vector data
   await deleteDatasetDataVector({ teamId, datasetIds });

packages/service/core/dataset/data/schema.ts

@@ -77,21 +77,32 @@ const DatasetDataSchema = new Schema({
   rebuilding: Boolean
 });
 
-// list collection and count data; list data; delete collection(relate data)
-DatasetDataSchema.index({
-  teamId: 1,
-  datasetId: 1,
-  collectionId: 1,
-  chunkIndex: 1,
-  updateTime: -1
+DatasetDataSchema.virtual('collection', {
+  ref: DatasetColCollectionName,
+  localField: 'collectionId',
+  foreignField: '_id',
+  justOne: true
 });
-// full text index
-DatasetDataSchema.index({ teamId: 1, datasetId: 1, fullTextToken: 'text' });
-// Recall vectors after data matching
-DatasetDataSchema.index({ teamId: 1, datasetId: 1, collectionId: 1, 'indexes.dataId': 1 });
-DatasetDataSchema.index({ updateTime: 1 });
-// rebuild data
-DatasetDataSchema.index({ rebuilding: 1, teamId: 1, datasetId: 1 });
+
+try {
+  // list collection and count data; list data; delete collection(relate data)
+  DatasetDataSchema.index({
+    teamId: 1,
+    datasetId: 1,
+    collectionId: 1,
+    chunkIndex: 1,
+    updateTime: -1
+  });
+  // full text index
+  DatasetDataSchema.index({ teamId: 1, datasetId: 1, fullTextToken: 'text' });
+  // Recall vectors after data matching
+  DatasetDataSchema.index({ teamId: 1, datasetId: 1, collectionId: 1, 'indexes.dataId': 1 });
+  DatasetDataSchema.index({ updateTime: 1 });
+  // rebuild data
+  DatasetDataSchema.index({ rebuilding: 1, teamId: 1, datasetId: 1 });
+} catch (error) {
+  console.log(error);
+}
 
 export const MongoDatasetData = getMongoModel<DatasetDataSchemaType>(
   DatasetDataCollectionName,

packages/service/core/dataset/search/controller.ts

@@ -8,8 +8,8 @@ import { getVectorsByText } from '../../ai/embedding';
 import { getVectorModel } from '../../ai/model';
 import { MongoDatasetData } from '../data/schema';
 import {
+  DatasetCollectionSchemaType,
   DatasetDataSchemaType,
-  DatasetDataWithCollectionType,
   SearchDataResponseItemType
 } from '@fastgpt/global/core/dataset/type';
 import { MongoDatasetCollection } from '../collection/schema';
@@ -267,7 +267,7 @@ export async function searchDatasetData(props: SearchDatasetDataProps) {
     });
 
     // get q and a
-    const dataList = (await MongoDatasetData.find(
+    const dataList = await MongoDatasetData.find(
       {
         teamId,
         datasetId: { $in: datasetIds },
@@ -276,8 +276,11 @@ export async function searchDatasetData(props: SearchDatasetDataProps) {
       },
       'datasetId collectionId updateTime q a chunkIndex indexes'
     )
-      .populate('collectionId', 'name fileId rawLink externalFileId externalFileUrl')
-      .lean()) as DatasetDataWithCollectionType[];
+      .populate<{ collection: DatasetCollectionSchemaType }>(
+        'collection',
+        'name fileId rawLink externalFileId externalFileUrl'
+      )
+      .lean();
 
     // add score to data(It's already sorted. The first one is the one with the most points)
     const concatResults = dataList.map((data) => {
@@ -307,8 +310,8 @@ export async function searchDatasetData(props: SearchDatasetDataProps) {
         a: data.a,
         chunkIndex: data.chunkIndex,
         datasetId: String(data.datasetId),
-        collectionId: String(data.collectionId?._id),
-        ...getCollectionSourceData(data.collectionId),
+        collectionId: String(data.collectionId),
+        ...getCollectionSourceData(data.collection),
         score: [{ type: SearchScoreTypeEnum.embedding, value: data.score, index }]
       };
 

packages/service/core/dataset/training/controller.ts

@@ -34,7 +34,7 @@ export const pushDataListToTrainingQueueByCollectionId = async ({
   session?: ClientSession;
 } & PushDatasetDataProps) => {
   const {
-    datasetId: { _id: datasetId, agentModel, vectorModel }
+    dataset: { _id: datasetId, agentModel, vectorModel }
   } = await getCollectionWithDataset(collectionId);
   return pushDataListToTrainingQueue({
     ...props,

packages/service/core/plugin/type.d.ts

@@ -2,6 +2,5 @@ import { PluginTemplateType } from '@fastgpt/global/core/plugin/type.d';
 import { SystemPluginTemplateItemType } from '@fastgpt/global/core/workflow/type';
 
 declare global {
-  var communityPluginsV1: PluginTemplateType[];
   var communityPlugins: SystemPluginTemplateItemType[];
 }

packages/service/core/workflow/dispatch/agent/classifyQuestion.ts

@@ -1,5 +1,8 @@
 import { chats2GPTMessages } from '@fastgpt/global/core/chat/adapt';
-import { countMessagesTokens } from '../../../../common/string/tiktoken/index';
+import {
+  countGptMessagesTokens,
+  countPromptTokens
+} from '../../../../common/string/tiktoken/index';
 import type { ChatItemType } from '@fastgpt/global/core/chat/type.d';
 import { ChatItemValueTypeEnum, ChatRoleEnum } from '@fastgpt/global/core/chat/constants';
 import { createChatCompletion } from '../../../ai/config';
@@ -35,7 +38,7 @@ type ActionProps = Props & { cqModel: LLMModelItemType };
 /* request openai chat */
 export const dispatchClassifyQuestion = async (props: Props): Promise<CQResponse> => {
   const {
-    user,
+    externalProvider,
     node: { nodeId, name },
     histories,
     params: { model, history = 6, agents, userChatInput }
@@ -49,7 +52,7 @@ export const dispatchClassifyQuestion = async (props: Props): Promise<CQResponse
 
   const chatHistories = getHistories(history, histories);
 
-  const { arg, tokens } = await completions({
+  const { arg, inputTokens, outputTokens } = await completions({
     ...props,
     histories: chatHistories,
     cqModel
@@ -59,7 +62,8 @@ export const dispatchClassifyQuestion = async (props: Props): Promise<CQResponse
 
   const { totalPoints, modelName } = formatModelChars2Points({
     model: cqModel.model,
-    tokens,
+    inputTokens: inputTokens,
+    outputTokens: outputTokens,
     modelType: ModelTypeEnum.llm
   });
 
@@ -69,10 +73,11 @@ export const dispatchClassifyQuestion = async (props: Props): Promise<CQResponse
       .filter((item) => item.key !== result.key)
       .map((item) => getHandleId(nodeId, 'source', item.key)),
     [DispatchNodeResponseKeyEnum.nodeResponse]: {
-      totalPoints: user.openaiAccount?.key ? 0 : totalPoints,
+      totalPoints: externalProvider.openaiAccount?.key ? 0 : totalPoints,
       model: modelName,
       query: userChatInput,
-      tokens,
+      inputTokens: inputTokens,
+      outputTokens: outputTokens,
       cqList: agents,
       cqResult: result.value,
       contextTotalLen: chatHistories.length + 2
@@ -80,9 +85,10 @@ export const dispatchClassifyQuestion = async (props: Props): Promise<CQResponse
     [DispatchNodeResponseKeyEnum.nodeDispatchUsages]: [
       {
         moduleName: name,
-        totalPoints: user.openaiAccount?.key ? 0 : totalPoints,
+        totalPoints: externalProvider.openaiAccount?.key ? 0 : totalPoints,
         model: modelName,
-        tokens
+        inputTokens: inputTokens,
+        outputTokens: outputTokens
       }
     ]
   };
@@ -90,7 +96,7 @@ export const dispatchClassifyQuestion = async (props: Props): Promise<CQResponse
 
 const completions = async ({
   cqModel,
-  user,
+  externalProvider,
   histories,
   params: { agents, systemPrompt = '', userChatInput }
 }: ActionProps) => {
@@ -131,7 +137,7 @@ const completions = async ({
       },
       cqModel
     ),
-    userKey: user.openaiAccount
+    userKey: externalProvider.openaiAccount
   });
   const answer = data.choices?.[0].message?.content || '';
 
@@ -148,7 +154,8 @@ const completions = async ({
   }
 
   return {
-    tokens: await countMessagesTokens(messages),
+    inputTokens: await countGptMessagesTokens(requestMessages),
+    outputTokens: await countPromptTokens(answer),
     arg: { type: id }
   };
 };

packages/service/core/workflow/dispatch/agent/extract.ts

@@ -3,7 +3,8 @@ import { filterGPTMessageByMaxTokens, loadRequestMessages } from '../../../chat/
 import type { ChatItemType } from '@fastgpt/global/core/chat/type.d';
 import {
   countMessagesTokens,
-  countGptMessagesTokens
+  countGptMessagesTokens,
+  countPromptTokens
 } from '../../../../common/string/tiktoken/index';
 import { ChatItemValueTypeEnum, ChatRoleEnum } from '@fastgpt/global/core/chat/constants';
 import { createChatCompletion } from '../../../ai/config';
@@ -46,7 +47,7 @@ const agentFunName = 'request_function';
 
 export async function dispatchContentExtract(props: Props): Promise<Response> {
   const {
-    user,
+    externalProvider,
     node: { name },
     histories,
     params: { content, history = 6, model, description, extractKeys }
@@ -59,7 +60,7 @@ export async function dispatchContentExtract(props: Props): Promise<Response> {
   const extractModel = getLLMModel(model);
   const chatHistories = getHistories(history, histories);
 
-  const { arg, tokens } = await (async () => {
+  const { arg, inputTokens, outputTokens } = await (async () => {
     if (extractModel.toolChoice) {
       return toolChoice({
         ...props,
@@ -114,7 +115,8 @@ export async function dispatchContentExtract(props: Props): Promise<Response> {
 
   const { totalPoints, modelName } = formatModelChars2Points({
     model: extractModel.model,
-    tokens,
+    inputTokens: inputTokens,
+    outputTokens: outputTokens,
     modelType: ModelTypeEnum.llm
   });
 
@@ -123,10 +125,11 @@ export async function dispatchContentExtract(props: Props): Promise<Response> {
     [NodeOutputKeyEnum.contextExtractFields]: JSON.stringify(arg),
     ...arg,
     [DispatchNodeResponseKeyEnum.nodeResponse]: {
-      totalPoints: user.openaiAccount?.key ? 0 : totalPoints,
+      totalPoints: externalProvider.openaiAccount?.key ? 0 : totalPoints,
       model: modelName,
       query: content,
-      tokens,
+      inputTokens,
+      outputTokens,
       extractDescription: description,
       extractResult: arg,
       contextTotalLen: chatHistories.length + 2
@@ -134,9 +137,10 @@ export async function dispatchContentExtract(props: Props): Promise<Response> {
     [DispatchNodeResponseKeyEnum.nodeDispatchUsages]: [
       {
         moduleName: name,
-        totalPoints: user.openaiAccount?.key ? 0 : totalPoints,
+        totalPoints: externalProvider.openaiAccount?.key ? 0 : totalPoints,
         model: modelName,
-        tokens
+        inputTokens,
+        outputTokens
       }
     ]
   };
@@ -211,7 +215,7 @@ ${description ? `- ${description}` : ''}
 };
 
 const toolChoice = async (props: ActionProps) => {
-  const { user, extractModel } = props;
+  const { externalProvider, extractModel } = props;
 
   const { filterMessages, agentFunction } = await getFunctionCallSchema(props);
 
@@ -233,7 +237,7 @@ const toolChoice = async (props: ActionProps) => {
       },
       extractModel
     ),
-    userKey: user.openaiAccount
+    userKey: externalProvider.openaiAccount
   });
 
   const arg: Record<string, any> = (() => {
@@ -249,21 +253,24 @@ const toolChoice = async (props: ActionProps) => {
     }
   })();
 
-  const completeMessages: ChatCompletionMessageParam[] = [
-    ...filterMessages,
+  const AIMessages: ChatCompletionMessageParam[] = [
     {
       role: ChatCompletionRequestMessageRoleEnum.Assistant,
       tool_calls: response.choices?.[0]?.message?.tool_calls
     }
   ];
+
+  const inputTokens = await countGptMessagesTokens(filterMessages, tools);
+  const outputTokens = await countGptMessagesTokens(AIMessages);
   return {
-    tokens: await countGptMessagesTokens(completeMessages, tools),
+    inputTokens,
+    outputTokens,
     arg
   };
 };
 
 const functionCall = async (props: ActionProps) => {
-  const { user, extractModel } = props;
+  const { externalProvider, extractModel } = props;
 
   const { agentFunction, filterMessages } = await getFunctionCallSchema(props);
   const functions: ChatCompletionCreateParams.Function[] = [agentFunction];
@@ -281,22 +288,26 @@ const functionCall = async (props: ActionProps) => {
       },
       extractModel
     ),
-    userKey: user.openaiAccount
+    userKey: externalProvider.openaiAccount
   });
 
   try {
     const arg = JSON.parse(response?.choices?.[0]?.message?.function_call?.arguments || '');
-    const completeMessages: ChatCompletionMessageParam[] = [
-      ...filterMessages,
+
+    const AIMessages: ChatCompletionMessageParam[] = [
       {
         role: ChatCompletionRequestMessageRoleEnum.Assistant,
         function_call: response.choices?.[0]?.message?.function_call
       }
     ];
 
+    const inputTokens = await countGptMessagesTokens(filterMessages, undefined, functions);
+    const outputTokens = await countGptMessagesTokens(AIMessages);
+
     return {
       arg,
-      tokens: await countGptMessagesTokens(completeMessages, undefined, functions)
+      inputTokens,
+      outputTokens
     };
   } catch (error) {
     console.log(response.choices?.[0]?.message);
@@ -305,14 +316,15 @@ const functionCall = async (props: ActionProps) => {
 
     return {
       arg: {},
-      tokens: 0
+      inputTokens: 0,
+      outputTokens: 0
     };
   }
 };
 
 const completions = async ({
   extractModel,
-  user,
+  externalProvider,
   histories,
   params: { content, extractKeys, description = 'No special requirements' }
 }: ActionProps) => {
@@ -360,7 +372,7 @@ Human: ${content}`
       },
       extractModel
     ),
-    userKey: user.openaiAccount
+    userKey: externalProvider.openaiAccount
   });
   const answer = data.choices?.[0].message?.content || '';
 
@@ -370,7 +382,8 @@ Human: ${content}`
   if (!jsonStr) {
     return {
       rawResponse: answer,
-      tokens: await countMessagesTokens(messages),
+      inputTokens: await countMessagesTokens(messages),
+      outputTokens: await countPromptTokens(answer),
       arg: {}
     };
   }
@@ -378,7 +391,8 @@ Human: ${content}`
   try {
     return {
       rawResponse: answer,
-      tokens: await countMessagesTokens(messages),
+      inputTokens: await countMessagesTokens(messages),
+      outputTokens: await countPromptTokens(answer),
       arg: json5.parse(jsonStr) as Record<string, any>
     };
   } catch (error) {
@@ -386,7 +400,8 @@ Human: ${content}`
     console.log(error);
     return {
       rawResponse: answer,
-      tokens: await countMessagesTokens(messages),
+      inputTokens: await countMessagesTokens(messages),
+      outputTokens: await countPromptTokens(answer),
       arg: {}
     };
   }

packages/service/core/workflow/dispatch/agent/runTool/functionCall.ts

@@ -43,10 +43,10 @@ export const runToolWithFunctionCall = async (
     requestOrigin,
     runtimeNodes,
     runtimeEdges,
-    user,
+    externalProvider,
     stream,
     workflowStreamResponse,
-    params: { temperature = 0, maxToken = 4000, aiChatVision }
+    params: { temperature, maxToken, aiChatVision }
   } = workflowProps;
 
   // Interactive
@@ -109,7 +109,8 @@ export const runToolWithFunctionCall = async (
 
       return {
         dispatchFlowResponse: [toolRunResponse],
-        toolNodeTokens: 0,
+        toolNodeInputTokens: 0,
+        toolNodeOutputTokens: 0,
         completeMessages: requestMessages,
         assistantResponses: toolRunResponse.assistantResponses,
         runTimes: toolRunResponse.runTimes,
@@ -126,7 +127,8 @@ export const runToolWithFunctionCall = async (
       },
       {
         dispatchFlowResponse: [toolRunResponse],
-        toolNodeTokens: 0,
+        toolNodeInputTokens: 0,
+        toolNodeOutputTokens: 0,
         assistantResponses: toolRunResponse.assistantResponses,
         runTimes: toolRunResponse.runTimes
       }
@@ -221,7 +223,7 @@ export const runToolWithFunctionCall = async (
     getEmptyResponseTip
   } = await createChatCompletion({
     body: requestBody,
-    userKey: user.openaiAccount,
+    userKey: externalProvider.openaiAccount,
     options: {
       headers: {
         Accept: 'application/json, text/plain, */*'
@@ -340,7 +342,9 @@ export const runToolWithFunctionCall = async (
       assistantToolMsgParams
     ] as ChatCompletionMessageParam[];
     // Only toolCall tokens are counted here, Tool response tokens count towards the next reply
-    const tokens = await countGptMessagesTokens(concatToolMessages, undefined, functions);
+    // const tokens = await countGptMessagesTokens(concatToolMessages, undefined, functions);
+    const inputTokens = await countGptMessagesTokens(requestMessages, undefined, functions);
+    const outputTokens = await countGptMessagesTokens([assistantToolMsgParams]);
     /* 
       ...
       user
@@ -375,7 +379,12 @@ export const runToolWithFunctionCall = async (
     const runTimes =
       (response?.runTimes || 0) +
       flatToolsResponseData.reduce((sum, item) => sum + item.runTimes, 0);
-    const toolNodeTokens = response?.toolNodeTokens ? response.toolNodeTokens + tokens : tokens;
+    const toolNodeInputTokens = response?.toolNodeInputTokens
+      ? response.toolNodeInputTokens + inputTokens
+      : inputTokens;
+    const toolNodeOutputTokens = response?.toolNodeOutputTokens
+      ? response.toolNodeOutputTokens + outputTokens
+      : outputTokens;
 
     // Check stop signal
     const hasStopSignal = flatToolsResponseData.some(
@@ -408,7 +417,8 @@ export const runToolWithFunctionCall = async (
 
       return {
         dispatchFlowResponse,
-        toolNodeTokens,
+        toolNodeInputTokens,
+        toolNodeOutputTokens,
         completeMessages,
         assistantResponses: toolNodeAssistants,
         runTimes,
@@ -423,7 +433,8 @@ export const runToolWithFunctionCall = async (
       },
       {
         dispatchFlowResponse,
-        toolNodeTokens,
+        toolNodeInputTokens,
+        toolNodeOutputTokens,
         assistantResponses: toolNodeAssistants,
         runTimes
       }
@@ -435,7 +446,8 @@ export const runToolWithFunctionCall = async (
       content: answer
     };
     const completeMessages = filterMessages.concat(gptAssistantResponse);
-    const tokens = await countGptMessagesTokens(completeMessages, undefined, functions);
+    const inputTokens = await countGptMessagesTokens(requestMessages, undefined, functions);
+    const outputTokens = await countGptMessagesTokens([gptAssistantResponse]);
     // console.log(tokens, 'response token');
 
     // concat tool assistant
@@ -443,7 +455,12 @@ export const runToolWithFunctionCall = async (
 
     return {
       dispatchFlowResponse: response?.dispatchFlowResponse || [],
-      toolNodeTokens: response?.toolNodeTokens ? response.toolNodeTokens + tokens : tokens,
+      toolNodeInputTokens: response?.toolNodeInputTokens
+        ? response.toolNodeInputTokens + inputTokens
+        : inputTokens,
+      toolNodeOutputTokens: response?.toolNodeOutputTokens
+        ? response.toolNodeOutputTokens + outputTokens
+        : outputTokens,
       completeMessages,
       assistantResponses: [...assistantResponses, ...toolNodeAssistant.value],
       runTimes: (response?.runTimes || 0) + 1

packages/service/core/workflow/dispatch/agent/runTool/index.ts

@@ -46,7 +46,7 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
     requestOrigin,
     chatConfig,
     runningAppInfo: { teamId },
-    user,
+    externalProvider,
     params: {
       model,
       systemPrompt,
@@ -153,7 +153,7 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
   })();
 
   // censor model and system key
-  if (toolModel.censor && !user.openaiAccount?.key) {
+  if (toolModel.censor && !externalProvider.openaiAccount?.key) {
     await postTextCensor({
       text: `${systemPrompt}
           ${userChatInput}
@@ -165,6 +165,8 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
     toolWorkflowInteractiveResponse,
     dispatchFlowResponse, // tool flow response
     toolNodeTokens,
+    toolNodeInputTokens,
+    toolNodeOutputTokens,
     completeMessages = [], // The actual message sent to AI(just save text)
     assistantResponses = [], // FastGPT system store assistant.value response
     runTimes
@@ -225,10 +227,11 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
 
   const { totalPoints, modelName } = formatModelChars2Points({
     model,
-    tokens: toolNodeTokens,
+    inputTokens: toolNodeInputTokens,
+    outputTokens: toolNodeOutputTokens,
     modelType: ModelTypeEnum.llm
   });
-  const toolAIUsage = user.openaiAccount?.key ? 0 : totalPoints;
+  const toolAIUsage = externalProvider.openaiAccount?.key ? 0 : totalPoints;
 
   // flat child tool response
   const childToolResponse = dispatchFlowResponse.map((item) => item.flowResponses).flat();
@@ -255,6 +258,8 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
       // 展示的积分消耗
       totalPoints: totalPointsUsage,
       toolCallTokens: toolNodeTokens,
+      toolCallInputTokens: toolNodeInputTokens,
+      toolCallOutputTokens: toolNodeOutputTokens,
       childTotalPoints: flatUsages.reduce((sum, item) => sum + item.totalPoints, 0),
       model: modelName,
       query: userChatInput,
@@ -270,9 +275,10 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
       // 工具调用本身的积分消耗
       {
         moduleName: name,
-        totalPoints: toolAIUsage,
         model: modelName,
-        tokens: toolNodeTokens
+        totalPoints: toolAIUsage,
+        inputTokens: toolNodeInputTokens,
+        outputTokens: toolNodeOutputTokens
       },
       // 工具的消耗
       ...flatUsages

packages/service/core/workflow/dispatch/agent/runTool/promptCall.ts

@@ -51,10 +51,10 @@ export const runToolWithPromptCall = async (
     requestOrigin,
     runtimeNodes,
     runtimeEdges,
-    user,
+    externalProvider,
     stream,
     workflowStreamResponse,
-    params: { temperature = 0, maxToken = 4000, aiChatVision }
+    params: { temperature, maxToken, aiChatVision }
   } = workflowProps;
 
   if (interactiveEntryToolParams) {
@@ -115,7 +115,8 @@ export const runToolWithPromptCall = async (
 
       return {
         dispatchFlowResponse: [toolRunResponse],
-        toolNodeTokens: 0,
+        toolNodeInputTokens: 0,
+        toolNodeOutputTokens: 0,
         completeMessages: concatMessages,
         assistantResponses: toolRunResponse.assistantResponses,
         runTimes: toolRunResponse.runTimes,
@@ -131,7 +132,8 @@ export const runToolWithPromptCall = async (
       },
       {
         dispatchFlowResponse: [toolRunResponse],
-        toolNodeTokens: 0,
+        toolNodeInputTokens: 0,
+        toolNodeOutputTokens: 0,
         assistantResponses: toolRunResponse.assistantResponses,
         runTimes: toolRunResponse.runTimes
       }
@@ -230,7 +232,7 @@ export const runToolWithPromptCall = async (
     getEmptyResponseTip
   } = await createChatCompletion({
     body: requestBody,
-    userKey: user.openaiAccount,
+    userKey: externalProvider.openaiAccount,
     options: {
       headers: {
         Accept: 'application/json, text/plain, */*'
@@ -286,15 +288,20 @@ export const runToolWithPromptCall = async (
       content: replaceAnswer
     };
     const completeMessages = filterMessages.concat(gptAssistantResponse);
-    const tokens = await countGptMessagesTokens(completeMessages, undefined);
-    // console.log(tokens, 'response token');
+    const inputTokens = await countGptMessagesTokens(requestMessages);
+    const outputTokens = await countGptMessagesTokens([gptAssistantResponse]);
 
     // concat tool assistant
     const toolNodeAssistant = GPTMessages2Chats([gptAssistantResponse])[0] as AIChatItemType;
 
     return {
       dispatchFlowResponse: response?.dispatchFlowResponse || [],
-      toolNodeTokens: response?.toolNodeTokens ? response.toolNodeTokens + tokens : tokens,
+      toolNodeInputTokens: response?.toolNodeInputTokens
+        ? response.toolNodeInputTokens + inputTokens
+        : inputTokens,
+      toolNodeOutputTokens: response?.toolNodeOutputTokens
+        ? response.toolNodeOutputTokens + outputTokens
+        : outputTokens,
       completeMessages,
       assistantResponses: [...assistantResponses, ...toolNodeAssistant.value],
       runTimes: (response?.runTimes || 0) + 1
@@ -366,17 +373,9 @@ export const runToolWithPromptCall = async (
     function_call: toolJson
   };
 
-  /* 
-    ...
-    user
-    assistant: tool data
-  */
-  const concatToolMessages = [
-    ...requestMessages,
-    assistantToolMsgParams
-  ] as ChatCompletionMessageParam[];
   // Only toolCall tokens are counted here, Tool response tokens count towards the next reply
-  const tokens = await countGptMessagesTokens(concatToolMessages, undefined);
+  const inputTokens = await countGptMessagesTokens(requestMessages);
+  const outputTokens = await countGptMessagesTokens([assistantToolMsgParams]);
 
   /* 
     ...
@@ -437,7 +436,12 @@ ANSWER: `;
   }
 
   const runTimes = (response?.runTimes || 0) + toolsRunResponse.toolResponse.runTimes;
-  const toolNodeTokens = response?.toolNodeTokens ? response.toolNodeTokens + tokens : tokens;
+  const toolNodeInputTokens = response?.toolNodeInputTokens
+    ? response.toolNodeInputTokens + inputTokens
+    : inputTokens;
+  const toolNodeOutputTokens = response?.toolNodeOutputTokens
+    ? response.toolNodeOutputTokens + outputTokens
+    : outputTokens;
 
   // Check stop signal
   const hasStopSignal = toolsRunResponse.toolResponse.flowResponses.some((item) => !!item.toolStop);
@@ -460,7 +464,8 @@ ANSWER: `;
 
     return {
       dispatchFlowResponse,
-      toolNodeTokens,
+      toolNodeInputTokens,
+      toolNodeOutputTokens,
       completeMessages: filterMessages,
       assistantResponses: toolNodeAssistants,
       runTimes,
@@ -475,7 +480,8 @@ ANSWER: `;
     },
     {
       dispatchFlowResponse,
-      toolNodeTokens,
+      toolNodeInputTokens,
+      toolNodeOutputTokens,
       assistantResponses: toolNodeAssistants,
       runTimes
     }

packages/service/core/workflow/dispatch/agent/runTool/toolChoice.ts

@@ -24,14 +24,13 @@ import { AIChatItemType } from '@fastgpt/global/core/chat/type';
 import { formatToolResponse, initToolCallEdges, initToolNodes } from './utils';
 import { computedMaxToken, llmCompletionsBodyFormat } from '../../../../ai/utils';
 import { getNanoid, sliceStrStartEnd } from '@fastgpt/global/common/string/tools';
-import { addLog } from '../../../../../common/system/log';
 import { toolValueTypeList } from '@fastgpt/global/core/workflow/constants';
 import { WorkflowInteractiveResponseType } from '@fastgpt/global/core/workflow/template/system/interactive/type';
 import { ChatItemValueTypeEnum } from '@fastgpt/global/core/chat/constants';
-import { i18nT } from '../../../../../../web/i18n/utils';
+import { getErrText } from '@fastgpt/global/common/error/utils';
 
 type ToolRunResponseType = {
-  toolRunResponse: DispatchFlowResponse;
+  toolRunResponse?: DispatchFlowResponse;
   toolMsgParams: ChatCompletionToolMessageParam;
 }[];
 
@@ -92,9 +91,9 @@ export const runToolWithToolChoice = async (
     runtimeNodes,
     runtimeEdges,
     stream,
-    user,
+    externalProvider,
     workflowStreamResponse,
-    params: { temperature = 0, maxToken = 4000, aiChatVision }
+    params: { temperature, maxToken, aiChatVision }
   } = workflowProps;
 
   if (maxRunToolTimes <= 0 && response) {
@@ -160,7 +159,8 @@ export const runToolWithToolChoice = async (
 
       return {
         dispatchFlowResponse: [toolRunResponse],
-        toolNodeTokens: 0,
+        toolNodeInputTokens: 0,
+        toolNodeOutputTokens: 0,
         completeMessages: requestMessages,
         assistantResponses: toolRunResponse.assistantResponses,
         runTimes: toolRunResponse.runTimes,
@@ -178,7 +178,8 @@ export const runToolWithToolChoice = async (
       },
       {
         dispatchFlowResponse: [toolRunResponse],
-        toolNodeTokens: 0,
+        toolNodeInputTokens: 0,
+        toolNodeOutputTokens: 0,
         assistantResponses: toolRunResponse.assistantResponses,
         runTimes: toolRunResponse.runTimes
       }
@@ -278,7 +279,7 @@ export const runToolWithToolChoice = async (
     getEmptyResponseTip
   } = await createChatCompletion({
     body: requestBody,
-    userKey: user.openaiAccount,
+    userKey: externalProvider.openaiAccount,
     options: {
       headers: {
         Accept: 'application/json, text/plain, */*'
@@ -344,59 +345,87 @@ export const runToolWithToolChoice = async (
     return Promise.reject(getEmptyResponseTip());
   }
 
-  // Run the selected tool by LLM.
-  const toolsRunResponse = (
-    await Promise.all(
-      toolCalls.map(async (tool) => {
-        const toolNode = toolNodes.find((item) => item.nodeId === tool.function?.name);
+  /* Run the selected tool by LLM.
+    Since only reference parameters are passed, if the same tool is run in parallel, it will get the same run parameters
+  */
+  const toolsRunResponse: ToolRunResponseType = [];
+  for await (const tool of toolCalls) {
+    try {
+      const toolNode = toolNodes.find((item) => item.nodeId === tool.function?.name);
 
-        if (!toolNode) return;
+      if (!toolNode) continue;
 
-        const startParams = (() => {
-          try {
-            return json5.parse(tool.function.arguments);
-          } catch (error) {
-            return {};
+      const startParams = (() => {
+        try {
+          return json5.parse(tool.function.arguments);
+        } catch (error) {
+          return {};
+        }
+      })();
+
+      initToolNodes(runtimeNodes, [toolNode.nodeId], startParams);
+      const toolRunResponse = await dispatchWorkFlow({
+        ...workflowProps,
+        isToolCall: true
+      });
+
+      const stringToolResponse = formatToolResponse(toolRunResponse.toolResponses);
+
+      const toolMsgParams: ChatCompletionToolMessageParam = {
+        tool_call_id: tool.id,
+        role: ChatCompletionRequestMessageRoleEnum.Tool,
+        name: tool.function.name,
+        content: stringToolResponse
+      };
+
+      workflowStreamResponse?.({
+        event: SseResponseEventEnum.toolResponse,
+        data: {
+          tool: {
+            id: tool.id,
+            toolName: '',
+            toolAvatar: '',
+            params: '',
+            response: sliceStrStartEnd(stringToolResponse, 5000, 5000)
           }
-        })();
+        }
+      });
 
-        initToolNodes(runtimeNodes, [toolNode.nodeId], startParams);
-        const toolRunResponse = await dispatchWorkFlow({
-          ...workflowProps,
-          isToolCall: true
-        });
+      toolsRunResponse.push({
+        toolRunResponse,
+        toolMsgParams
+      });
+    } catch (error) {
+      const err = getErrText(error);
+      workflowStreamResponse?.({
+        event: SseResponseEventEnum.toolResponse,
+        data: {
+          tool: {
+            id: tool.id,
+            toolName: '',
+            toolAvatar: '',
+            params: '',
+            response: sliceStrStartEnd(err, 5000, 5000)
+          }
+        }
+      });
 
-        const stringToolResponse = formatToolResponse(toolRunResponse.toolResponses);
-
-        const toolMsgParams: ChatCompletionToolMessageParam = {
+      toolsRunResponse.push({
+        toolRunResponse: undefined,
+        toolMsgParams: {
           tool_call_id: tool.id,
           role: ChatCompletionRequestMessageRoleEnum.Tool,
           name: tool.function.name,
-          content: stringToolResponse
-        };
+          content: sliceStrStartEnd(err, 5000, 5000)
+        }
+      });
+    }
+  }
 
-        workflowStreamResponse?.({
-          event: SseResponseEventEnum.toolResponse,
-          data: {
-            tool: {
-              id: tool.id,
-              toolName: '',
-              toolAvatar: '',
-              params: '',
-              response: sliceStrStartEnd(stringToolResponse, 5000, 5000)
-            }
-          }
-        });
-
-        return {
-          toolRunResponse,
-          toolMsgParams
-        };
-      })
-    )
-  ).filter(Boolean) as ToolRunResponseType;
-
-  const flatToolsResponseData = toolsRunResponse.map((item) => item.toolRunResponse).flat();
+  const flatToolsResponseData = toolsRunResponse
+    .map((item) => item.toolRunResponse)
+    .flat()
+    .filter(Boolean) as DispatchFlowResponse[];
   // concat tool responses
   const dispatchFlowResponse = response
     ? response.dispatchFlowResponse.concat(flatToolsResponseData)
@@ -430,24 +459,26 @@ export const runToolWithToolChoice = async (
     ] as ChatCompletionMessageParam[];
 
     // Only toolCall tokens are counted here, Tool response tokens count towards the next reply
-    const tokens = await countGptMessagesTokens(concatToolMessages, tools);
+    const inputTokens = await countGptMessagesTokens(requestMessages, tools);
+    const outputTokens = await countGptMessagesTokens(assistantToolMsgParams);
+
     /* 
-        ...
-        user
-        assistant: tool data
-        tool: tool response
-      */
+      ...
+      user
+      assistant: tool data
+      tool: tool response
+    */
     const completeMessages = [
       ...concatToolMessages,
       ...toolsRunResponse.map((item) => item?.toolMsgParams)
     ];
 
     /* 
-        Get tool node assistant response
-        history assistant
-        current tool assistant
-        tool child assistant
-      */
+      Get tool node assistant response
+      history assistant
+      current tool assistant
+      tool child assistant
+    */
     const toolNodeAssistant = GPTMessages2Chats([
       ...assistantToolMsgParams,
       ...toolsRunResponse.map((item) => item?.toolMsgParams)
@@ -465,7 +496,10 @@ export const runToolWithToolChoice = async (
     const runTimes =
       (response?.runTimes || 0) +
       flatToolsResponseData.reduce((sum, item) => sum + item.runTimes, 0);
-    const toolNodeTokens = response ? response.toolNodeTokens + tokens : tokens;
+    const toolNodeInputTokens = response ? response.toolNodeInputTokens + inputTokens : inputTokens;
+    const toolNodeOutputTokens = response
+      ? response.toolNodeOutputTokens + outputTokens
+      : outputTokens;
 
     // Check stop signal
     const hasStopSignal = flatToolsResponseData.some(
@@ -473,12 +507,12 @@ export const runToolWithToolChoice = async (
     );
     // Check interactive response(Only 1 interaction is reserved)
     const workflowInteractiveResponseItem = toolsRunResponse.find(
-      (item) => item.toolRunResponse.workflowInteractiveResponse
+      (item) => item.toolRunResponse?.workflowInteractiveResponse
     );
     if (hasStopSignal || workflowInteractiveResponseItem) {
       // Get interactive tool data
       const workflowInteractiveResponse =
-        workflowInteractiveResponseItem?.toolRunResponse.workflowInteractiveResponse;
+        workflowInteractiveResponseItem?.toolRunResponse?.workflowInteractiveResponse;
 
       // Flashback traverses completeMessages, intercepting messages that know the first user
       const firstUserIndex = completeMessages.findLastIndex((item) => item.role === 'user');
@@ -498,7 +532,8 @@ export const runToolWithToolChoice = async (
 
       return {
         dispatchFlowResponse,
-        toolNodeTokens,
+        toolNodeInputTokens,
+        toolNodeOutputTokens,
         completeMessages,
         assistantResponses: toolNodeAssistants,
         runTimes,
@@ -514,7 +549,8 @@ export const runToolWithToolChoice = async (
       },
       {
         dispatchFlowResponse,
-        toolNodeTokens,
+        toolNodeInputTokens,
+        toolNodeOutputTokens,
         assistantResponses: toolNodeAssistants,
         runTimes
       }
@@ -526,14 +562,17 @@ export const runToolWithToolChoice = async (
       content: answer
     };
     const completeMessages = filterMessages.concat(gptAssistantResponse);
-    const tokens = await countGptMessagesTokens(completeMessages, tools);
+    const inputTokens = await countGptMessagesTokens(requestMessages, tools);
+    const outputTokens = await countGptMessagesTokens([gptAssistantResponse]);
 
     // concat tool assistant
     const toolNodeAssistant = GPTMessages2Chats([gptAssistantResponse])[0] as AIChatItemType;
 
     return {
       dispatchFlowResponse: response?.dispatchFlowResponse || [],
-      toolNodeTokens: response ? response.toolNodeTokens + tokens : tokens,
+      toolNodeInputTokens: response ? response.toolNodeInputTokens + inputTokens : inputTokens,
+      toolNodeOutputTokens: response ? response.toolNodeOutputTokens + outputTokens : outputTokens,
+
       completeMessages,
       assistantResponses: [...assistantResponses, ...toolNodeAssistant.value],
       runTimes: (response?.runTimes || 0) + 1
@@ -580,7 +619,8 @@ async function streamResponse({
           text: content
         })
       });
-    } else if (responseChoice?.tool_calls?.[0]) {
+    }
+    if (responseChoice?.tool_calls?.[0]) {
       const toolCall: ChatCompletionMessageToolCall = responseChoice.tool_calls[0];
       // In a stream response, only one tool is returned at a time.  If have id, description is executing a tool
       if (toolCall.id || callingTool) {

packages/service/core/workflow/dispatch/agent/runTool/type.d.ts

@@ -31,7 +31,9 @@ export type DispatchToolModuleProps = ModuleDispatchProps<{
 
 export type RunToolResponse = {
   dispatchFlowResponse: DispatchFlowResponse[];
-  toolNodeTokens: number;
+  toolNodeTokens?: number; // deprecated
+  toolNodeInputTokens: number;
+  toolNodeOutputTokens: number;
   completeMessages?: ChatCompletionMessageParam[];
   assistantResponses?: AIChatItemValueItemType[];
   toolWorkflowInteractiveResponse?: WorkflowInteractiveResponseType;

packages/service/core/workflow/dispatch/chat/oneapi.ts

@@ -5,13 +5,17 @@ import { ChatRoleEnum } from '@fastgpt/global/core/chat/constants';
 import { SseResponseEventEnum } from '@fastgpt/global/core/workflow/runtime/constants';
 import { textAdaptGptResponse } from '@fastgpt/global/core/workflow/runtime/utils';
 import { createChatCompletion } from '../../../ai/config';
-import type { ChatCompletion, StreamChatType } from '@fastgpt/global/core/ai/type.d';
+import type {
+  ChatCompletion,
+  ChatCompletionMessageParam,
+  StreamChatType
+} from '@fastgpt/global/core/ai/type.d';
 import { formatModelChars2Points } from '../../../../support/wallet/usage/utils';
 import type { LLMModelItemType } from '@fastgpt/global/core/ai/model.d';
 import { postTextCensor } from '../../../../common/api/requestPlusApi';
 import { ChatCompletionRequestMessageRoleEnum } from '@fastgpt/global/core/ai/constants';
 import type { DispatchNodeResultType } from '@fastgpt/global/core/workflow/runtime/type';
-import { countMessagesTokens } from '../../../../common/string/tiktoken/index';
+import { countGptMessagesTokens } from '../../../../common/string/tiktoken/index';
 import {
   chats2GPTMessages,
   chatValue2RuntimePrompt,
@@ -62,7 +66,7 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
     res,
     requestOrigin,
     stream = false,
-    user,
+    externalProvider,
     histories,
     node: { name },
     query,
@@ -71,8 +75,8 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
     chatConfig,
     params: {
       model,
-      temperature = 0,
-      maxToken = 4000,
+      temperature,
+      maxToken,
       history = 6,
       quoteQA,
       userChatInput,
@@ -134,7 +138,7 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
     }),
     (() => {
       // censor model and system key
-      if (modelConstantsData.censor && !user.openaiAccount?.key) {
+      if (modelConstantsData.censor && !externalProvider.openaiAccount?.key) {
         return postTextCensor({
           text: `${systemPrompt}
             ${userChatInput}
@@ -170,7 +174,7 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
   // console.log(JSON.stringify(requestBody, null, 2), '===');
   const { response, isStreamResponse, getEmptyResponseTip } = await createChatCompletion({
     body: requestBody,
-    userKey: user.openaiAccount,
+    userKey: externalProvider.openaiAccount,
     options: {
       headers: {
         Accept: 'application/json, text/plain, */*'
@@ -214,25 +218,34 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
     return Promise.reject(getEmptyResponseTip());
   }
 
-  const completeMessages = requestMessages.concat({
-    role: ChatCompletionRequestMessageRoleEnum.Assistant,
-    content: answerText
-  });
+  const AIMessages: ChatCompletionMessageParam[] = [
+    {
+      role: ChatCompletionRequestMessageRoleEnum.Assistant,
+      content: answerText
+    }
+  ];
+
+  const completeMessages = [...requestMessages, ...AIMessages];
   const chatCompleteMessages = GPTMessages2Chats(completeMessages);
 
-  const tokens = await countMessagesTokens(chatCompleteMessages);
+  const inputTokens = await countGptMessagesTokens(requestMessages);
+  const outputTokens = await countGptMessagesTokens(AIMessages);
+
   const { totalPoints, modelName } = formatModelChars2Points({
     model,
-    tokens,
+    inputTokens,
+    outputTokens,
     modelType: ModelTypeEnum.llm
   });
 
   return {
     answerText,
     [DispatchNodeResponseKeyEnum.nodeResponse]: {
-      totalPoints: user.openaiAccount?.key ? 0 : totalPoints,
+      totalPoints: externalProvider.openaiAccount?.key ? 0 : totalPoints,
       model: modelName,
-      tokens,
+      tokens: inputTokens + outputTokens,
+      inputTokens: inputTokens,
+      outputTokens: outputTokens,
       query: `${userChatInput}`,
       maxToken: max_tokens,
       historyPreview: getHistoryPreview(
@@ -245,9 +258,10 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
     [DispatchNodeResponseKeyEnum.nodeDispatchUsages]: [
       {
         moduleName: name,
-        totalPoints: user.openaiAccount?.key ? 0 : totalPoints,
+        totalPoints: externalProvider.openaiAccount?.key ? 0 : totalPoints,
         model: modelName,
-        tokens
+        inputTokens: inputTokens,
+        outputTokens: outputTokens
       }
     ],
     [DispatchNodeResponseKeyEnum.toolResponses]: answerText,

packages/service/core/workflow/dispatch/dataset/search.ts

@@ -120,14 +120,14 @@ export async function dispatchDatasetSearch(
   // vector
   const { totalPoints, modelName } = formatModelChars2Points({
     model: vectorModel.model,
-    tokens,
+    inputTokens: tokens,
     modelType: ModelTypeEnum.vector
   });
   const responseData: DispatchNodeResponseType & { totalPoints: number } = {
     totalPoints,
     query: concatQueries.join('\n'),
     model: modelName,
-    tokens,
+    inputTokens: tokens,
     similarity: usingSimilarityFilter ? similarity : undefined,
     limit,
     searchMode,
@@ -139,19 +139,21 @@ export async function dispatchDatasetSearch(
       totalPoints,
       moduleName: node.name,
       model: modelName,
-      tokens
+      inputTokens: tokens
     }
   ];
 
   if (aiExtensionResult) {
     const { totalPoints, modelName } = formatModelChars2Points({
       model: aiExtensionResult.model,
-      tokens: aiExtensionResult.tokens,
+      inputTokens: aiExtensionResult.inputTokens,
+      outputTokens: aiExtensionResult.outputTokens,
       modelType: ModelTypeEnum.llm
     });
 
     responseData.totalPoints += totalPoints;
-    responseData.tokens = aiExtensionResult.tokens;
+    responseData.inputTokens = aiExtensionResult.inputTokens;
+    responseData.outputTokens = aiExtensionResult.outputTokens;
     responseData.extensionModel = modelName;
     responseData.extensionResult =
       aiExtensionResult.extensionQueries?.join('\n') ||
@@ -161,7 +163,8 @@ export async function dispatchDatasetSearch(
       totalPoints,
       moduleName: 'core.module.template.Query extension',
       model: modelName,
-      tokens: aiExtensionResult.tokens
+      inputTokens: aiExtensionResult.inputTokens,
+      outputTokens: aiExtensionResult.outputTokens
     });
   }