Skip to main content
Glama
Thread Analyzer MCP Server

Thread Analyzer MCP Server


What is this?

An MCP server that lets your AI assistant scrape a Threads post URL, then query and analyze the replies — all through natural conversation.

Instead of:

1. Manually open browser
2. Scroll through hundreds of replies
3. Copy-paste into spreadsheet
4. Manually look for patterns

Just say:

"Analyze the replies on this Threads post: https://www.threads.com/@zuck/post/ABC123"

Your AI agent handles the rest.

Related MCP server: Social Fetch MCP

Features

  • Network interception — Captures GraphQL API responses, not fragile CSS selectors that Meta randomizes

  • Anti-detection — Randomized scroll delays, stealth browser flags, custom User-Agent

  • 4 MCP tools — Scrape, list, search, and get statistics on replies

  • Works with any MCP client — Claude Code, Claude Desktop, Cursor, VS Code, Windsurf, and more

MCP Tools

Tool

Description

scrape_thread(url)

Scrape all replies from a public Threads post

get_all_replies()

Return all scraped replies with username and timestamp

search_replies(keyword)

Case-insensitive keyword search across replies

get_reply_stats()

Reply count, top commenters, avg length, time range

Quick Start

Prerequisites

  • Python 3.13+

  • uv package manager

Installation

git clone https://github.com/ethan-tsai-tsai/thread-analyzer.git
cd thread-analyzer
uv sync
uv run playwright install chromium

Configuration

Add the server to your MCP client config:

{
  "mcpServers": {
    "thread-analyzer": {
      "command": "uv",
      "args": ["run", "--directory", "/absolute/path/to/thread-analyzer", "python", "server.py"]
    }
  }
}

Add to your project's .mcp.json:

{
  "mcpServers": {
    "thread-analyzer": {
      "command": "uv",
      "args": ["run", "--directory", "/absolute/path/to/thread-analyzer", "python", "server.py"]
    }
  }
}

Or run: claude mcp add thread-analyzer -- uv run --directory /absolute/path/to/thread-analyzer python server.py

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "thread-analyzer": {
      "command": "uv",
      "args": ["run", "--directory", "/absolute/path/to/thread-analyzer", "python", "server.py"]
    }
  }
}

Go to Cursor Settings > MCP > Add new MCP Server, then add:

{
  "mcpServers": {
    "thread-analyzer": {
      "command": "uv",
      "args": ["run", "--directory", "/absolute/path/to/thread-analyzer", "python", "server.py"]
    }
  }
}

Add to .vscode/mcp.json in your workspace:

{
  "servers": {
    "thread-analyzer": {
      "command": "uv",
      "args": ["run", "--directory", "/absolute/path/to/thread-analyzer", "python", "server.py"]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "thread-analyzer": {
      "command": "uv",
      "args": ["run", "--directory", "/absolute/path/to/thread-analyzer", "python", "server.py"]
    }
  }
}

Standalone CLI

You can also use the scraper directly without MCP:

uv run python scraper.py "https://www.threads.com/@user/post/XXXXX"

# Options
uv run python scraper.py "URL" --output custom.csv --max-scrolls 50

How It Works

┌─────────────┐     MCP (stdio)     ┌──────────────┐    Playwright    ┌─────────────┐
│  AI Client  │ ◄──────────────────► │  server.py   │ ◄──────────────► │  Threads.com │
│ (Claude,    │   scrape_thread()    │  (FastMCP)   │   GraphQL API   │  (Meta)      │
│  Cursor...) │   get_all_replies()  │              │   interception  │              │
│             │   search_replies()   │  replies.csv │                 │              │
│             │   get_reply_stats()  │              │                 │              │
└─────────────┘                      └──────────────┘                 └─────────────┘
  1. You give your AI assistant a Threads post URL

  2. AI calls scrape_thread(url) via MCP

  3. Server launches headless Chromium, navigates to the post

  4. Playwright intercepts GraphQL network responses containing reply data

  5. Server parses replies (username, text, timestamp), saves to CSV

  6. AI uses get_all_replies(), search_replies(), get_reply_stats() to analyze

Anti-Detection

Technique

Purpose

Custom User-Agent

Mimics real Chrome browser

navigator.webdriver removal

Hides automation flag

AutomationControlled disabled

Prevents Chromium detection

Randomized scroll delays (1.5-4.5s)

Avoids behavioral fingerprinting

Early stop on idle scrolls

Mimics natural browsing patterns

Limitations

  • Public posts only — Cannot access private or restricted posts

  • Meta's anti-bot measures — Meta may block headless browsers; if scraping fails, try the standalone CLI in non-headless mode

  • GraphQL schema changes — Meta periodically changes their API structure; the parser in scraper.py may need updating

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

MIT


Available Tools

4 tools
get_all_repliesA

Get all scraped Threads replies.

Returns a formatted list of all replies with username, timestamp and text. Call scrape_thread first if no data exists yet.

Inputs: no named parameters. Returns: result

get_reply_statsA

Get statistics about the scraped replies.

Returns reply count, unique users, most active commenters, and time range of replies.

Inputs: no named parameters. Returns: result

scrape_threadA

Scrape replies from a public Threads post URL.

Opens a browser, navigates to the post, intercepts GraphQL API responses to extract replies, and saves them locally. This must be called before using get_all_replies or search_replies.

Args: url: Full URL of the Threads post (e.g. https://www.threads.com/@user/post/ABC123) max_scrolls: Maximum scroll iterations for loading more replies (default: 20)

Inputs: url (required), max_scrolls. Returns: result

search_repliesA

Search Threads replies for a keyword (case-insensitive).

Args: keyword: The search term to filter replies by.

Returns matching replies with username and text.

Inputs: keyword (required). Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4.3/5.0

    Scored across 4 tools

    Disambiguation5/5

    Each tool has a distinct role: scrape_thread for fetching, get_all_replies for retrieving all, search_replies for filtering, get_reply_stats for analytics. No overlap or ambiguity.

    Naming Consistency5/5

    All tools follow a clear verb_noun snake_case pattern (scrape_, get_, search_), making the API predictable and easy to navigate.

    Tool Count5/5

    Four tools is well-scoped for a focused Threads analysis server, covering the full workflow without bloat or missing essentials.

    Completeness5/5

    The tool surface covers scraping, retrieval, search, and stats—everything needed for the stated purpose of analyzing Threads replies. No obvious gaps.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      MCP server for Threads comment research — analyze public sentiment, frustrations, and conversations on Threads via AI agents.
      MIT
    • A
      license
      C
      quality
      B
      maintenance
      Lets coding agents fetch real social media and web data from platforms like TikTok, Instagram, YouTube, and more, directly inside editors like Cursor and VS Code.
      2
      500
      1
      MIT
    • F
      license
      A
      quality
      D
      maintenance
      Enables AI agents to interact with Meta Threads by fetching profile details, recent posts, post and account analytics, and conversation reply trees, as well as exporting posts as JSON for archiving or embedding.
      7
      -
    • A
      license
      A
      quality
      B
      maintenance
      Enables AI agents to manage a Threads profile through Meta's API, including posting, threads, replies, reply moderation, and insights.
      30
      21 npm
      4
      MIT