meta-threads-mcp
Unofficial MCP server for Meta's Threads API. Enables LLMs like Claude to publish posts, manage replies, and track insights through the Model Context Protocol.
README
meta-threads-mcp
Unofficial MCP server for Meta's Threads API. Enables LLMs like Claude to publish posts, manage replies, and track insights through the Model Context Protocol.
Features
- Full Threads API coverage via MCP tools
- Create text, image, and video posts
- Reply management (reply, hide/unhide, get conversation)
- Post and user insights/metrics
- Publishing quota tracking
- Built on meta-threads-sdk and FastMCP
Installation
pip install meta-threads-mcp
Or with uv:
uv add meta-threads-mcp
Running on FastMCP Cloud
The easiest way to run this MCP server is to deploy it on FastMCP Cloud:
- Go to fastmcp.cloud and sign in
- Connect your GitHub repository
- FastMCP Cloud will automatically detect the
fastmcp.jsonconfiguration and deploy your server

Configuration
Claude Desktop
Add to your claude_desktop_config.json:
{
"mcpServers": {
"threads": {
"command": "meta-threads-mcp"
}
}
}
Authentication
The server expects a bearer token in the format:
<access_token>:<user_id>
The client passes this token via the request context when calling tools.
Available Tools
Posts
| Tool | Description | Parameters |
|---|---|---|
threads_create_post |
Create and publish a text post | text (required), reply_control (optional) |
threads_create_image_post |
Create post with image | image_url (required), text (optional) |
threads_create_video_post |
Create post with video | video_url (required), text (optional) |
threads_get_post |
Get a post by ID | post_id (required) |
threads_get_user_posts |
Get user's recent posts | limit (optional, default 10) |
threads_delete_post |
Delete a post | post_id (required) |
threads_get_publishing_limit |
Check remaining quota | none |
Replies
| Tool | Description | Parameters |
|---|---|---|
threads_reply_to_post |
Reply to a post | post_id (required), text (required) |
threads_get_replies |
Get replies to a post | post_id (required) |
threads_get_conversation |
Get full conversation thread | post_id (required) |
threads_hide_reply |
Hide a reply | reply_id (required) |
threads_unhide_reply |
Unhide a reply | reply_id (required) |
Insights
| Tool | Description | Parameters |
|---|---|---|
threads_get_media_insights |
Get metrics for a post | media_id (required) |
threads_get_user_insights |
Get user-level metrics | none |
User
| Tool | Description | Parameters |
|---|---|---|
threads_get_profile |
Get current user's profile | none |
Reply Control Options
When creating posts, you can control who can reply:
EVERYONE(default) - Anyone can replyACCOUNTS_YOU_FOLLOW- Only accounts you follow can replyMENTIONED_ONLY- Only mentioned accounts can reply
Rate Limits
- 250 posts per 24 hours
- 1000 replies per 24 hours
Use threads_get_publishing_limit to check your current quota.
Development
Setup
git clone https://github.com/MetaThreads/meta-threads-mcp.git
cd meta-threads-mcp
uv sync --dev
Running Tests
uv run pytest
Linting & Type Checking
uv run ruff check src tests
uv run ruff format src tests
uv run mypy src
Testing with FastMCP CLI
fastmcp dev src/meta_threads_mcp/server.py
License
MIT License - see LICENSE for details.
Related Projects
- meta-threads-sdk - Python SDK for Meta's Threads API
- FastMCP - Simple MCP server framework
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。