### Install Dependencies and Pull Environment Variables

Source: https://github.com/niapya/clawless/blob/main/README.md

Installs project dependencies and pulls Vercel environment variables locally for development. Ensure you have Bun installed.

```bash
cd your-clawless

bun install

# Pull Vercel environment variables locally, including AUTH_SECRET, USERNAME, PASSWORD, and KV, DB, Blob and Sandbox keys.
bun vercel pull

bun dev
```

--------------------------------

### GET /api/skills

Source: https://context7.com/niapya/clawless/llms.txt

Lists all installed skills.

```APIDOC
## GET /api/skills

### Description
Retrieves a list of all installed skills.

### Method
GET

### Endpoint
https://your-clawless.vercel.app/api/skills
```

--------------------------------

### List Skills

Source: https://context7.com/niapya/clawless/llms.txt

Retrieve metadata for all installed skills.

```bash
curl "https://your-clawless.vercel.app/api/skills" \
  -H "Cookie: clawless-auth=<token>"
```

```json
[
  {
    "name": "code-review",
    "description": "代码审查助手",
    "source": "git",
    "gitURL": "https://github.com/example/skills.git",
    "files": ["SKILL.md", "prompts/review.md"]
  },
  {
    "name": "translator",
    "description": "多语言翻译",
    "source": "manual",
    "files": ["SKILL.md"]
  }
]
```

--------------------------------

### Get Configuration

Source: https://context7.com/niapya/clawless/llms.txt

Retrieves the full application configuration, including AI models, channels, and agent settings.

```bash
# 获取当前配置
curl "https://your-clawless.vercel.app/api/config" \
  -H "Cookie: clawless-auth=<token>"

# 响应示例
{
  "config": {
    "models": {
      "model": "openai/gpt-4",
      "embedding_model": "openai/text-embedding-3-small",
      "temperature": 0.7,
      "providers": {
        "openai": {
          "format": "openaicompatible",
          "base_url": "https://api.openai.com/v1"
        }
      }
    },
    "channels": {
      "telegram": {
        "enabled": true,
        "bot_token": "xxx",
        "allowed_author_ids": ["user123"]
      }
    },
    "agents": {
      "default": {
        "system_prompt": "你是一个有帮助的 AI 助手",
        "temperature": 0.7
      }
    }
  },
  "runtimeHealth": {...}
}
```

--------------------------------

### Get Skill Details

Source: https://context7.com/niapya/clawless/llms.txt

Retrieve details for a specific skill.

```bash
curl "https://your-clawless.vercel.app/api/skills/my-skill" \
  -H "Cookie: clawless-auth=<token>"
```

```json
{
  "name": "my-skill",
  "description": "自定义技能",
  "source": "manual",
  "files": ["SKILL.md", "prompts/main.md"],
  "createdAt": "2024-01-01T00:00:00Z",
  "updatedAt": "2024-01-01T00:00:00Z"
}
```

--------------------------------

### GET /api/skills/{skillName}

Source: https://context7.com/niapya/clawless/llms.txt

Gets details for a specific skill.

```APIDOC
## GET /api/skills/{skillName}

### Description
Retrieves metadata for a specific skill.

### Method
GET

### Endpoint
https://your-clawless.vercel.app/api/skills/{skillName}

### Path Parameters
- **skillName** (string) - Required - Name of the skill
```

--------------------------------

### GET /api/files

Source: https://context7.com/niapya/clawless/llms.txt

Retrieves a list of uploaded files with optional filtering and pagination.

```APIDOC
## GET /api/files

### Description
Lists files uploaded to the system.

### Method
GET

### Endpoint
https://your-clawless.vercel.app/api/files

### Parameters
#### Query Parameters
- **limit** (integer) - Optional - Number of files to return
- **sort** (string) - Optional - Sort order
- **sessionId** (string) - Optional - Filter by session ID

### Response
#### Success Response (200)
- **files** (array) - List of file objects
- **hasMore** (boolean) - Pagination indicator

#### Response Example
{
  "files": [
    {
      "id": "file-123",
      "name": "document.pdf"
    }
  ],
  "hasMore": false
}
```

--------------------------------

### GET /api/memory/long-term

Source: https://context7.com/niapya/clawless/llms.txt

Searches or lists long-term memory entries.

```APIDOC
## GET /api/memory/long-term

### Description
Searches memory using semantic/keyword queries or lists all memories.

### Method
GET

### Endpoint
https://your-clawless.vercel.app/api/memory/long-term

### Query Parameters
- **query** (string) - Optional - Search query
- **minConfidence** (number) - Optional - Confidence threshold
- **pageSize** (number) - Optional - Page size
- **page** (number) - Optional - Page number
```

--------------------------------

### GET /api/schedules

Source: https://context7.com/niapya/clawless/llms.txt

Retrieves a list of all scheduled tasks, including daily and delayed tasks.

```APIDOC
## GET /api/schedules

### Description
Fetches all scheduled tasks.

### Method
GET

### Endpoint
https://your-clawless.vercel.app/api/schedules

### Response
#### Success Response (200)
- **tasks** (array) - List of scheduled task objects

#### Response Example
{
  "tasks": [
    {
      "id": "task-123",
      "type": "daily",
      "title": "每日摘要",
      "nextRunAt": "2024-01-02T01:00:00Z",
      "active": true
    }
  ]
}
```

--------------------------------

### Development and Deployment Commands

Source: https://github.com/niapya/clawless/blob/main/AGENTS.md

Common CLI commands for managing dependencies, database migrations, and deployment using Bun.

```bash
bun install
```

```bash
bun run db:generate  # 生成迁移
bun run db:push     # 推送到数据库
```

```bash
bun run dev
# 访问 http://localhost:3000
```

```bash
bun run check        # 完整检查
```

```bash
bun run build        # 构建生产版本
bun run deploy       # 部署到 Vercel
```

--------------------------------

### Replace Configuration via PUT

Source: https://context7.com/niapya/clawless/llms.txt

Perform a full replacement of the system configuration.

```bash
curl -X PUT "https://your-clawless.vercel.app/api/config/update" \
  -H "Content-Type: application/json" \
  -d '{
    "models": {
      "model": "openai/gpt-4",
      "embedding_model": "openai/text-embedding-3-small",
      "temperature": 0.7,
      "providers": {
        "openai": {
          "format": "openaicompatible",
          "api_key": "sk-xxx",
          "base_url": "https://api.openai.com/v1"
        }
      }
    },
    "channels": {
      "telegram": {
        "enabled": true,
        "bot_token": "your-bot-token",
        "allowed_author_ids": ["123456789"]
      }
    }
  }'
```

--------------------------------

### Create Skill Manually

Source: https://context7.com/niapya/clawless/llms.txt

Create a new skill by providing file paths and content.

```bash
curl -X POST "https://your-clawless.vercel.app/api/skills" \
  -H "Content-Type: application/json" \
  -H "Cookie: clawless-auth=<token>" \
  -d '{
    "name": "my-skill",
    "description": "自定义技能",
    "files": [
      {
        "path": "SKILL.md",
        "content": "# My Skill\n\n这是一个自定义技能的说明..."
      },
      {
        "path": "prompts/main.md",
        "content": "## 系统提示词\n\n你是一个..."
      }
    ]
  }'
```

```json
{
  "name": "my-skill",
  "description": "自定义技能",
  "source": "manual",
  "files": ["SKILL.md", "prompts/main.md"],
  "createdAt": "2024-01-01T00:00:00Z"
}
```

--------------------------------

### Import Skill from Git

Source: https://context7.com/niapya/clawless/llms.txt

Asynchronously import skills from a Git repository.

```bash
curl -X POST "https://your-clawless.vercel.app/api/skills/import" \
  -H "Content-Type: application/json" \
  -d '{
    "gitURL": "https://github.com/example/ai-skills.git"
  }'
```

```json
{
  "jobId": "job-uuid-123",
  "status": "pending",
  "gitURL": "https://github.com/example/ai-skills.git"
}
```

```bash
curl "https://your-clawless.vercel.app/api/skills/import?jobId=job-uuid-123"
```

```json
{
  "jobId": "job-uuid-123",
  "status": "syncing",
  "gitURL": "https://github.com/example/ai-skills.git"
}
```

```json
{
  "jobId": "job-uuid-123",
  "status": "done",
  "gitURL": "https://github.com/example/ai-skills.git",
  "importedNames": ["skill-a", "skill-b"],
  "removedNames": [],
  "fileCount": 12,
  "finishedAt": 1704067200000
}
```

```bash
curl "https://your-clawless.vercel.app/api/skills/import"
```

--------------------------------

### PUT /api/config/update

Source: https://context7.com/niapya/clawless/llms.txt

Completely replaces the system configuration.

```APIDOC
## PUT /api/config/update

### Description
Completely replaces the system configuration.

### Method
PUT

### Endpoint
https://your-clawless.vercel.app/api/config/update

### Request Body
- **models** (object) - Required - Model configuration
- **channels** (object) - Required - Channel configuration

### Request Example
{
  "models": {
    "model": "openai/gpt-4",
    "embedding_model": "openai/text-embedding-3-small",
    "temperature": 0.7,
    "providers": {
      "openai": {
        "format": "openaicompatible",
        "api_key": "sk-xxx",
        "base_url": "https://api.openai.com/v1"
      }
    }
  },
  "channels": {
    "telegram": {
      "enabled": true,
      "bot_token": "your-bot-token",
      "allowed_author_ids": ["123456789"]
    }
  }
}
```

--------------------------------

### Project Directory Structure

Source: https://github.com/niapya/clawless/blob/main/AGENTS.md

The standard directory layout for the Next.js App Router project.

```text
app/                     # Next.js App Router
├── (auth)/              # 认证相关路由
├── (chat)/              # 聊天相关路由
├── (config)/            # 配置相关路由
├── (memory)/            # 内存管理路由
├── (skill)/             # 技能管理路由
└── layout.tsx           # 根布局

components/              # React 可复用组件
├── ui/                  # shadcn/ui 原始组件
├── auth/                # 认证组件
├── config/              # 配置界面组件
└── *.tsx                # 其他功能组件

lib/                     # 工具函数与核心逻辑
├── ai/                  # AI SDK 封装
├── auth/                # 认证逻辑
├── blob/                # Blob 存储
├── bot/                 # Bot 适配器
├── chat/                # 聊天逻辑
├── db/                  # 数据库操作
├── kv/                  # Redis 缓存
├── sandbox/             # 沙箱执行
├── workflow/            # 工作流调度
└── utils/               # 通用工具函数

types/                   # TypeScript 类型定义
├── config/              # 配置相关类型
├── skills/              # 技能相关类型
└── *.ts                 # 其他类型

public/                  # 静态资源
└── fonts/, images/
```

--------------------------------

### POST /api/skills/import

Source: https://context7.com/niapya/clawless/llms.txt

Imports a skill from a Git repository.

```APIDOC
## POST /api/skills/import

### Description
Initiates an asynchronous import of a skill from a Git repository.

### Method
POST

### Endpoint
https://your-clawless.vercel.app/api/skills/import

### Request Body
- **gitURL** (string) - Required - URL of the Git repository
```

--------------------------------

### Subscribe to Workflow Stream

Source: https://context7.com/niapya/clawless/llms.txt

Establishes a connection to receive real-time streaming output from an active workflow.

```bash
# 订阅工作流输出流
curl "https://your-clawless.vercel.app/api/ai/{runId}/stream" \
  -H "Cookie: clawless-auth=<token>" \
  -H "Accept: text/event-stream"

# 响应: Server-Sent Events 流
# data: {"type":"text-delta","delta":"Hello"}
# data: {"type":"text-delta","delta":" World"}
```

--------------------------------

### POST /api/skills

Source: https://context7.com/niapya/clawless/llms.txt

Manually creates a new skill.

```APIDOC
## POST /api/skills

### Description
Creates a new skill by providing file content.

### Method
POST

### Endpoint
https://your-clawless.vercel.app/api/skills

### Request Body
- **name** (string) - Required - Skill name
- **description** (string) - Required - Skill description
- **files** (array) - Required - List of file objects
```

--------------------------------

### List Files

Source: https://context7.com/niapya/clawless/llms.txt

Retrieves a list of uploaded files with optional filtering by session ID.

```bash
curl "https://your-clawless.vercel.app/api/files?limit=30&sort=desc" \
  -H "Cookie: clawless-auth=<token>"
```

```bash
curl "https://your-clawless.vercel.app/api/files?sessionId=session-123&limit=20"
```

--------------------------------

### Update Configuration via PATCH

Source: https://context7.com/niapya/clawless/llms.txt

Perform a partial update of the system configuration.

```bash
curl -X PATCH "https://your-clawless.vercel.app/api/config/update" \
  -H "Content-Type: application/json" \
  -H "Cookie: clawless-auth=<token>" \
  -d '{
    "models": {
      "model": "anthropic/claude-3-5-sonnet-20241022",
      "temperature": 0.5
    }
  }'
```

--------------------------------

### Authenticate User

Source: https://context7.com/niapya/clawless/llms.txt

Logs in a user and returns an authentication cookie for subsequent requests.

```bash
# 用户登录
curl -X POST "https://your-clawless.vercel.app/api/auth/login" \
  -H "Content-Type: application/json" \
  -d '{
    "username": "admin",
    "password": "your-password"
  }'

# 响应示例
{
  "ok": true,
  "redirectTo": "/"
}
# Set-Cookie: clawless-auth=<token>; Path=/; HttpOnly; Secure; SameSite=Lax
```

--------------------------------

### Slash Commands

Source: https://context7.com/niapya/clawless/llms.txt

List of available slash commands for chat interactions.

```bash
/help                    # 显示帮助信息
/status                  # 显示当前会话状态
/new                     # 创建新会话
/session                 # 显示当前绑定的会话
/session <session-id>    # 切换到指定会话
/stop                    # 停止当前工作流
/compact                 # 请求上下文压缩
/approve <toolCallId>    # 批准待定的工具调用
/reject <toolCallId>     # 拒绝待定的工具调用
```

--------------------------------

### Search and List Long-Term Memory

Source: https://context7.com/niapya/clawless/llms.txt

Query memories using semantic search or list all existing entries.

```bash
curl "https://your-clawless.vercel.app/api/memory/long-term?query=TypeScript%20开发偏好&minConfidence=0.7&pageSize=10" \
  -H "Cookie: clawless-auth=<token>"
```

```json
{
  "search": true,
  "page": 1,
  "pageSize": 10,
  "results": [
    {
      "memoryId": "mem-123",
      "content": "用户偏好使用 TypeScript...",
      "confidence": 0.85,
      "matchType": "semantic"
    }
  ]
}
```

```bash
curl "https://your-clawless.vercel.app/api/memory/long-term?page=1&pageSize=50"
```

```json
{
  "search": false,
  "page": 1,
  "pageSize": 50,
  "items": [...]
}
```

--------------------------------

### TypeScript Import Order Convention

Source: https://github.com/niapya/clawless/blob/main/AGENTS.md

Standardized import ordering for maintaining clean and readable codebases.

```typescript
// 1. 外部库
import { useCallback, useEffect, useMemo, useRef, useState } from 'react';
import { useChat } from '@ai-sdk/react';
import { ofetch } from 'ofetch';

// 2. 内部组件与工具
import { ChatHeader } from '@/components/chat-header';
import { deriveSessionTitle } from '@/lib/chat/session-title';
import { type WorkflowDataPart } from '@/types/workflow';

// 3. 本地导入
import { Messages } from './messages';
```

--------------------------------

### Configure Telegram Webhook

Source: https://context7.com/niapya/clawless/llms.txt

Sets the webhook URL for Telegram bot integration.

```bash
curl -X POST "https://api.telegram.org/bot<BOT_TOKEN>/setWebhook" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://your-clawless.vercel.app/api/bot/<YOUR_SECRET>/telegram/callback",
    "secret_token": "<OPTIONAL_SECRET>"
  }'
```

--------------------------------

### AppConfig Schema

Source: https://context7.com/niapya/clawless/llms.txt

TypeScript interface defining the application configuration structure.

```typescript
interface AppConfig {
  models?: {
    model: string;                    // 格式: "provider/model-id"
    embedding_model?: string;         // 格式: "provider/model-id"
    temperature?: number;             // 0.0 - 2.0
    context_limit?: number;           // Token 上限
    max_output_tokens?: number;       // 输出 Token 上限
    providers?: Record<string, {
      format: 'openaicompatible' | 'anthropic' | 'openai' | 'google';
      api_key?: string;
      base_url?: string;
      headers?: Record<string, string>;
    }>;
  };

  agents?: Record<string, {
    model?: string;
    system_prompt?: string;
    temperature?: number;
  }>;

  channels?: {
    slack?: {
      enabled: boolean;
      bot_token?: string;
      signing_secret?: string;
      allowed_author_ids?: string[];
    };
    teams?: {
      enabled: boolean;
      app_id?: string;
      app_password?: string;
      allowed_author_ids?: string[];
    };
    gchat?: {
      enabled: boolean;
      project_id?: string;
      credentials_json?: string;
      allowed_author_ids?: string[];
    };
    telegram?: {
      enabled: boolean;
      bot_token?: string;
      secret_token?: string;
      bot_username?: string;
      allowed_author_ids?: string[];
    };
  };

  autonomy?: { /* 自主权限配置 */ };
  tools?: { /* 内置工具配置 */ };
  mcp?: { /* MCP 远程服务器配置 */ };
}
```

--------------------------------

### POST /api/memory/long-term

Source: https://context7.com/niapya/clawless/llms.txt

Creates a new long-term memory entry.

```APIDOC
## POST /api/memory/long-term

### Description
Creates a new long-term memory entry and generates vector embeddings.

### Method
POST

### Endpoint
https://your-clawless.vercel.app/api/memory/long-term

### Request Body
- **content** (string) - Required - The memory content

### Response
#### Success Response (200)
- **memory** (object) - Created memory object
- **indexing** (object) - Indexing metadata
```

--------------------------------

### Configuration Management API

Source: https://context7.com/niapya/clawless/llms.txt

Endpoints for retrieving and updating the application's configuration settings.

```APIDOC
## GET /api/config

### Description
Retrieves the complete application configuration, including settings for AI models, channels, agents, and tools.

### Method
GET

### Endpoint
/api/config

### Parameters
#### Request Headers
- **Cookie** (string) - Required - Authentication cookie (`clawless-auth=<token>`).

### Response
#### Success Response (200)
- **config** (object) - The application configuration object.
  - **models** (object) - AI model configurations.
    - **model** (string) - Default AI model.
    - **embedding_model** (string) - Default embedding model.
    - **temperature** (number) - Default temperature setting.
    - **providers** (object) - AI provider configurations.
      - **openai** (object) - OpenAI provider settings.
        - **format** (string) - API format compatibility.
        - **base_url** (string) - Base URL for the OpenAI API.
  - **channels** (object) - Channel integrations configuration.
    - **telegram** (object) - Telegram channel settings.
      - **enabled** (boolean) - Whether the Telegram channel is enabled.
      - **bot_token** (string) - The Telegram bot token.
      - **allowed_author_ids** (array) - List of allowed author IDs.
  - **agents** (object) - Agent configurations.
    - **default** (object) - Default agent settings.
      - **system_prompt** (string) - The system prompt for the default agent.
      - **temperature** (number) - Temperature setting for the default agent.
- **runtimeHealth** (object) - Information about the runtime health of the application.

#### Response Example
```json
{
  "config": {
    "models": {
      "model": "openai/gpt-4",
      "embedding_model": "openai/text-embedding-3-small",
      "temperature": 0.7,
      "providers": {
        "openai": {
          "format": "openaicompatible",
          "base_url": "https://api.openai.com/v1"
        }
      }
    },
    "channels": {
      "telegram": {
        "enabled": true,
        "bot_token": "xxx",
        "allowed_author_ids": ["user123"]
      }
    },
    "agents": {
      "default": {
        "system_prompt": "你是一个有帮助的 AI 助手",
        "temperature": 0.7
      }
    }
  },
  "runtimeHealth": {...}
}
```
```

```APIDOC
## PATCH /api/config or PUT /api/config

### Description
Partially updates (PATCH) or completely replaces (PUT) the application configuration.

### Method
PATCH or PUT

### Endpoint
/api/config

### Parameters
#### Request Body
- **config** (object) - The configuration object to update or replace. Structure is the same as the response of `GET /api/config`.

#### Request Headers
- **Cookie** (string) - Required - Authentication cookie (`clawless-auth=<token>`).
- **Content-Type** (string) - Required - `application/json` for PUT, or `application/json-patch+json` for PATCH.

### Response
#### Success Response (200)
- **ok** (boolean) - Indicates if the configuration update was successful.

#### Response Example
```json
{ "ok": true }
```
```

--------------------------------

### Manage Sessions

Source: https://context7.com/niapya/clawless/llms.txt

Lists sessions or retrieves message history for a specific session with pagination support.

```bash
# 列出所有会话
curl "https://your-clawless.vercel.app/api/sessions?limit=50&channel=web" \
  -H "Cookie: clawless-auth=<token>"

# 响应
{
  "sessions": [
    {
      "id": "session-123",
      "title": "Python 脚本编写",
      "channel": "web",
      "status": "active",
      "totalTokens": 2500,
      "createdAt": "2024-01-01T00:00:00Z"
    }
  ]
}

# 获取指定会话的消息（带分页）
curl "https://your-clawless.vercel.app/api/sessions?id=session-123&limit=20&before=2024-01-01T12:00:00Z"

# 响应
{
  "sessionId": "session-123",
  "messages": [...],
  "hasMore": true,
  "nextBefore": "2024-01-01T10:00:00Z"
}
```

--------------------------------

### Create Long-Term Memory

Source: https://context7.com/niapya/clawless/llms.txt

Create a new memory entry with automatic vector embedding generation.

```bash
curl -X POST "https://your-clawless.vercel.app/api/memory/long-term" \
  -H "Content-Type: application/json" \
  -H "Cookie: clawless-auth=<token>" \
  -d '{
    "content": "用户偏好使用 TypeScript 进行开发，喜欢函数式编程风格"
  }'
```

```json
{
  "memory": {
    "id": "mem-123",
    "content": "用户偏好使用 TypeScript 进行开发...",
    "createdAt": "2024-01-01T00:00:00Z",
    "updatedAt": "2024-01-01T00:00:00Z"
  },
  "indexing": {
    "mode": "embedded",
    "embeddingModel": "openai/text-embedding-3-small",
    "embeddingDimensions": 1536,
    "warning": null
  }
}
```

--------------------------------

### List Scheduled Tasks

Source: https://context7.com/niapya/clawless/llms.txt

Retrieves a list of all scheduled tasks, including daily and delayed types.

```bash
curl "https://your-clawless.vercel.app/api/schedules" \
  -H "Cookie: clawless-auth=<token>"
```

--------------------------------

### Interact with Running Workflow

Source: https://context7.com/niapya/clawless/llms.txt

Sends user messages or control commands to an active workflow execution.

```bash
# 发送用户消息
curl -X POST "https://your-clawless.vercel.app/api/ai/{runId}/message" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "user-message",
    "message": "继续执行下一步",
    "uiMessageId": "msg-789"
  }'

# 发送控制命令 (compact/cancel)
curl -X POST "https://your-clawless.vercel.app/api/ai/{runId}/message" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "control",
    "command": "compact",
    "reason": "上下文过长"
  }'

# 响应
{ "ok": true }
```

--------------------------------

### PUT /api/schedules/{taskId}

Source: https://context7.com/niapya/clawless/llms.txt

Updates the configuration of an existing scheduled task.

```APIDOC
## PUT /api/schedules/{taskId}

### Description
Updates a specific scheduled task by ID.

### Method
PUT

### Endpoint
https://your-clawless.vercel.app/api/schedules/{taskId}

### Parameters
#### Path Parameters
- **taskId** (string) - Required - The ID of the task to update

#### Request Body
- **type** (string) - Required - Task type (daily or delay)
- **title** (string) - Optional - Task title
- **prompt** (string) - Optional - Task prompt
- **active** (boolean) - Optional - Task status

### Response
#### Success Response (200)
- **task** (object) - The updated task object

#### Response Example
{
  "task": {
    "id": "task-123",
    "type": "daily",
    "title": "每日报告"
  }
}
```

--------------------------------

### Delete Skill

Source: https://context7.com/niapya/clawless/llms.txt

Remove a skill and its associated files.

```bash
curl -X DELETE "https://your-clawless.vercel.app/api/skills/my-skill" \
  -H "Cookie: clawless-auth=<token>"
```

--------------------------------

### Manage AI Chat Messages

Source: https://context7.com/niapya/clawless/llms.txt

Handles sending new messages or regenerating existing ones using the main AI endpoint.

```bash
# 发送新消息
curl -X POST "https://your-clawless.vercel.app/api/ai" \
  -H "Content-Type: application/json" \
  -H "Cookie: clawless-auth=<token>" \
  -d '{
    "id": "session-uuid-here",
    "trigger": "submit-message",
    "messageId": "msg-uuid-here",
    "input": {
      "text": "你好，请帮我写一个 Python 脚本",
      "parts": [
        { "type": "text", "text": "你好，请帮我写一个 Python 脚本" }
      ]
    }
  }'

# 重新生成消息
curl -X POST "https://your-clawless.vercel.app/api/ai" \
  -H "Content-Type: application/json" \
  -d '{
    "id": "session-uuid",
    "trigger": "regenerate-message",
    "messageId": "last-message-id"
  }'

# 响应: Server-Sent Events 流
# Headers: x-session-id, x-workflow-run-id
```

--------------------------------

### PATCH /api/config/update

Source: https://context7.com/niapya/clawless/llms.txt

Partially updates the system configuration, such as model settings.

```APIDOC
## PATCH /api/config/update

### Description
Partially updates the system configuration.

### Method
PATCH

### Endpoint
https://your-clawless.vercel.app/api/config/update

### Request Body
- **models** (object) - Optional - Configuration for models

### Request Example
{
  "models": {
    "model": "anthropic/claude-3-5-sonnet-20241022",
    "temperature": 0.5
  }
}
```

--------------------------------

### Update Scheduled Tasks

Source: https://context7.com/niapya/clawless/llms.txt

Updates the configuration for a specific task. Supports both delay and daily task types.

```bash
curl -X PUT "https://your-clawless.vercel.app/api/schedules/{taskId}" \
  -H "Content-Type: application/json" \
  -H "Cookie: clawless-auth=<token>" \
  -d '{
    "type": "delay",
    "title": "延迟提醒",
    "prompt": "提醒用户完成代码审查",
    "runAt": "2024-01-01T15:00:00Z",
    "active": true
  }'
```

```bash
curl -X PUT "https://your-clawless.vercel.app/api/schedules/{taskId}" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "daily",
    "title": "每日报告",
    "prompt": "生成并发送每日工作报告",
    "dailyTime": "18:00",
    "timezone": "Asia/Shanghai",
    "active": true
  }'
```

--------------------------------

### Query Workflow Status

Source: https://context7.com/niapya/clawless/llms.txt

Retrieves the current status, session details, and token usage for a specific workflow run.

```bash
# 获取运行状态
curl "https://your-clawless.vercel.app/api/ai/{runId}/status" \
  -H "Cookie: clawless-auth=<token>"

# 响应示例
{
  "runId": "run-123",
  "sessionId": "session-456",
  "status": "running",
  "workflow": {
    "phase": "running",
    "lastRunId": "run-123",
    "startedAt": "2024-01-01T00:00:00Z"
  },
  "session": {
    "id": "session-456",
    "channel": "web",
    "model": "openai/gpt-4",
    "totalTokens": 1500,
    "source": { "type": "web" }
  }
}
```

--------------------------------

### User Authentication API

Source: https://context7.com/niapya/clawless/llms.txt

Endpoints for user authentication, primarily for obtaining authentication tokens.

```APIDOC
## POST /api/auth/login

### Description
Logs in a user by verifying username and password, returning an authentication cookie for subsequent requests.

### Method
POST

### Endpoint
/api/auth/login

### Parameters
#### Request Body
- **username** (string) - Required - The username for login.
- **password** (string) - Required - The password for login.

### Request Example
```json
{
  "username": "admin",
  "password": "your-password"
}
```

### Response
#### Success Response (200)
- **ok** (boolean) - Indicates if the login was successful.
- **redirectTo** (string) - The URL to redirect to after successful login.

#### Response Example
```json
{
  "ok": true,
  "redirectTo": "/"
}
```

**Set-Cookie**: `clawless-auth=<token>; Path=/; HttpOnly; Secure; SameSite=Lax`
```

--------------------------------

### Session Management API

Source: https://context7.com/niapya/clawless/llms.txt

Endpoints for managing user sessions, including listing, retrieving messages, and deleting sessions.

```APIDOC
## GET /api/sessions

### Description
Retrieves a list of sessions or the message history for a specific session. Supports pagination and channel filtering.

### Method
GET

### Endpoint
/api/sessions

### Parameters
#### Query Parameters
- **limit** (integer) - Optional - The maximum number of sessions or messages to return.
- **channel** (string) - Optional - Filters sessions by channel (e.g., "web").
- **id** (string) - Optional - Retrieves messages for a specific session ID.
- **before** (string) - Optional - Retrieves messages before a specific timestamp (ISO 8601 format), used for pagination.

#### Request Headers
- **Cookie** (string) - Required - Authentication cookie (`clawless-auth=<token>`).

### Response (List Sessions)
#### Success Response (200)
- **sessions** (array) - An array of session objects.
  - **id** (string) - The unique ID of the session.
  - **title** (string) - The title or summary of the session.
  - **channel** (string) - The communication channel.
  - **status** (string) - The current status of the session (e.g., "active").
  - **totalTokens** (integer) - The total tokens used in the session.
  - **createdAt** (string) - The timestamp when the session was created.

#### Response Example (List Sessions)
```json
{
  "sessions": [
    {
      "id": "session-123",
      "title": "Python 脚本编写",
      "channel": "web",
      "status": "active",
      "totalTokens": 2500,
      "createdAt": "2024-01-01T00:00:00Z"
    }
  ]
}
```

### Response (Get Session Messages)
#### Success Response (200)
- **sessionId** (string) - The ID of the session.
- **messages** (array) - An array of message objects.
- **hasMore** (boolean) - Indicates if there are more messages available.
- **nextBefore** (string) - The timestamp to use for fetching the next page of messages.

#### Response Example (Get Session Messages)
```json
{
  "sessionId": "session-123",
  "messages": [...],
  "hasMore": true,
  "nextBefore": "2024-01-01T10:00:00Z"
}
```
```

```APIDOC
## DELETE /api/sessions/{sessionId}

### Description
Deletes a specific session, along with any associated workflows and sandbox processes.

### Method
DELETE

### Endpoint
/api/sessions/{sessionId}

### Parameters
#### Path Parameters
- **sessionId** (string) - Required - The ID of the session to delete.

#### Request Headers
- **Cookie** (string) - Required - Authentication cookie (`clawless-auth=<token>`).

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the deletion was successful.

#### Response Example
```json
{ "success": true }
```
```

--------------------------------

### Update Long-Term Memory

Source: https://context7.com/niapya/clawless/llms.txt

Update memory content and trigger re-embedding.

```bash
curl -X PATCH "https://your-clawless.vercel.app/api/memory/long-term/{memoryId}" \
  -H "Content-Type: application/json" \
  -d '{
    "content": "用户偏好使用 TypeScript 和 React 进行前端开发"
  }'
```

```json
{
  "memory": {...},
  "indexing": {...}
}
```

--------------------------------

### DELETE /api/skills/{skillName}

Source: https://context7.com/niapya/clawless/llms.txt

Deletes a specific skill from the system.

```APIDOC
## DELETE /api/skills/{skillName}

### Description
Deletes a specific skill by its name.

### Method
DELETE

### Endpoint
https://your-clawless.vercel.app/api/skills/{skillName}

### Parameters
#### Path Parameters
- **skillName** (string) - Required - The name of the skill to delete

### Response
#### Success Response (200)
- **success** (boolean) - Status of the operation
- **name** (string) - Name of the deleted skill

#### Response Example
{
  "success": true,
  "name": "my-skill"
}
```

--------------------------------

### PATCH /api/memory/long-term/{memoryId}

Source: https://context7.com/niapya/clawless/llms.txt

Updates an existing memory entry.

```APIDOC
## PATCH /api/memory/long-term/{memoryId}

### Description
Updates the content of a specific memory entry.

### Method
PATCH

### Endpoint
https://your-clawless.vercel.app/api/memory/long-term/{memoryId}

### Path Parameters
- **memoryId** (string) - Required - ID of the memory to update
```

--------------------------------

### Delete Session

Source: https://context7.com/niapya/clawless/llms.txt

Removes a session and terminates its associated workflow and sandbox.

```bash
# 删除会话
curl -X DELETE "https://your-clawless.vercel.app/api/sessions/{sessionId}" \
  -H "Cookie: clawless-auth=<token>"

# 响应
{ "success": true }
```

--------------------------------

### AI Chat API

Source: https://context7.com/niapya/clawless/llms.txt

Endpoints for interacting with the AI chat functionality, including sending messages, managing workflow status, and streaming responses.

```APIDOC
## POST /api/ai

### Description
Main chat endpoint to send new messages, regenerate messages, or route messages. Returns a streaming response in UI Message Stream format.

### Method
POST

### Endpoint
/api/ai

### Parameters
#### Request Body
- **id** (string) - Required - The session ID.
- **trigger** (string) - Required - The trigger for the action (e.g., "submit-message", "regenerate-message").
- **messageId** (string) - Required - The ID of the message to interact with.
- **input** (object) - Required for "submit-message" trigger - Contains the user's input.
  - **text** (string) - Required - The text content of the message.
  - **parts** (array) - Required - An array of message parts.
    - **type** (string) - Required - The type of the part (e.g., "text").
    - **text** (string) - Required - The text content of the part.

### Request Example
```json
{
  "id": "session-uuid-here",
  "trigger": "submit-message",
  "messageId": "msg-uuid-here",
  "input": {
    "text": "你好，请帮我写一个 Python 脚本",
    "parts": [
      { "type": "text", "text": "你好，请帮我写一个 Python 脚本" }
    ]
  }
}
```

### Response
#### Success Response (200)
Returns a Server-Sent Events (SSE) stream.

**Headers**: `x-session-id`, `x-workflow-run-id`
```

```APIDOC
## GET /api/ai/{runId}/status

### Description
Retrieves the status information for a current workflow run, including session details, run status, and token usage.

### Method
GET

### Endpoint
/api/ai/{runId}/status

### Parameters
#### Path Parameters
- **runId** (string) - Required - The ID of the workflow run.

#### Request Headers
- **Cookie** (string) - Required - Authentication cookie (`clawless-auth=<token>`).

### Response
#### Success Response (200)
- **runId** (string) - The ID of the workflow run.
- **sessionId** (string) - The ID of the session.
- **status** (string) - The current status of the run (e.g., "running").
- **workflow** (object) - Workflow-specific details.
  - **phase** (string) - The current phase of the workflow.
  - **lastRunId** (string) - The ID of the last workflow run.
  - **startedAt** (string) - The timestamp when the workflow started.
- **session** (object) - Session-specific details.
  - **id** (string) - The ID of the session.
  - **channel** (string) - The communication channel used.
  - **model** (string) - The AI model being used.
  - **totalTokens** (integer) - The total number of tokens used.
  - **source** (object) - Information about the session source.
    - **type** (string) - The type of the source (e.g., "web").

#### Response Example
```json
{
  "runId": "run-123",
  "sessionId": "session-456",
  "status": "running",
  "workflow": {
    "phase": "running",
    "lastRunId": "run-123",
    "startedAt": "2024-01-01T00:00:00Z"
  },
  "session": {
    "id": "session-456",
    "channel": "web",
    "model": "openai/gpt-4",
    "totalTokens": 1500,
    "source": { "type": "web" }
  }
}
```
```

```APIDOC
## GET /api/ai/{runId}/stream

### Description
Subscribes to the real-time stream of output from an active workflow run.

### Method
GET

### Endpoint
/api/ai/{runId}/stream

### Parameters
#### Path Parameters
- **runId** (string) - Required - The ID of the workflow run.

#### Request Headers
- **Cookie** (string) - Required - Authentication cookie (`clawless-auth=<token>`).
- **Accept** (string) - Required - Must be `text/event-stream`.

### Response
#### Success Response (200)
Returns a Server-Sent Events (SSE) stream.

**Example Stream Data**:
```
data: {"type":"text-delta","delta":"Hello"}
data: {"type":"text-delta","delta":" World"}
```
```

```APIDOC
## POST /api/ai/{runId}/message

### Description
Sends a user message, system message, or control command to an active workflow run.

### Method
POST

### Endpoint
/api/ai/{runId}/message

### Parameters
#### Path Parameters
- **runId** (string) - Required - The ID of the workflow run.

#### Request Body
- **type** (string) - Required - The type of message (e.g., "user-message", "control").
- **message** (string) - Required if type is "user-message" - The user's message content.
- **uiMessageId** (string) - Required if type is "user-message" - The UI message ID.
- **command** (string) - Required if type is "control" - The control command (e.g., "compact", "cancel").
- **reason** (string) - Optional - The reason for the control command.

### Request Example (User Message)
```json
{
  "type": "user-message",
  "message": "继续执行下一步",
  "uiMessageId": "msg-789"
}
```

### Request Example (Control Command)
```json
{
  "type": "control",
  "command": "compact",
  "reason": "上下文过长"
}
```

### Response
#### Success Response (200)
- **ok** (boolean) - Indicates if the operation was successful.

#### Response Example
```json
{ "ok": true }
```
```

```APIDOC
## POST /api/ai/{runId}/pause

### Description
Pauses or cancels a running workflow.

### Method
POST

### Endpoint
/api/ai/{runId}/pause

### Parameters
#### Path Parameters
- **runId** (string) - Required - The ID of the workflow run to pause.

#### Request Headers
- **Cookie** (string) - Required - Authentication cookie (`clawless-auth=<token>`).

### Response
#### Success Response (200)
- **ok** (boolean) - Indicates if the operation was successful.
- **sessionId** (string) - The ID of the session.
- **runId** (string) - The ID of the workflow run that was paused.

#### Response Example
```json
{
  "ok": true,
  "sessionId": "session-456",
  "runId": "run-123"
}
```
```

--------------------------------

### Pause Workflow

Source: https://context7.com/niapya/clawless/llms.txt

Stops an active workflow execution and cancels the current process.

```bash
# 暂停/取消工作流
curl -X POST "https://your-clawless.vercel.app/api/ai/{runId}/pause" \
  -H "Cookie: clawless-auth=<token>"

# 响应
{
  "ok": true,
  "sessionId": "session-456",
  "runId": "run-123"
}
```