Skip to main content
Glama
fetcher-sh

fetcher.sh MCP Server

Official
by fetcher-sh

fetcher.sh MCP Server

npm version MIT License MCP Spec 2025-11-25

Official Model Context Protocol server for fetcher.sh — web data API for AI agents and applications.

Connect Claude Desktop, Cursor, VS Code Copilot, or any MCP client to fetcher.sh and search tweets, YouTube videos, Reddit threads, Google results, app-store listings, and more — all from natural language.

Features

Five catalog tools, mapped to the hosted server at mcp.fetcher.sh. Discovery is local (no network); paid fetches hit the live REST API.

Tool

Cost

What it does

search_endpoints

free

Keyword / service search over 111 endpoints across 11 APIs

describe_endpoint

free

Parameters, allowed values, price, response shape

fetch_data

paid

Run any catalog path: { path, params }

check_balance

free

Credits left on FETCHER_API_KEY

topup_credits

Returns the prepaid top-up URL (minimum $1)

Set FETCHER_SERVICE=twitter (or youtube, reddit, …) to switch to a specialized server: one named tool per endpoint of that service, plus the two credit tools.

Related MCP server: entroute

Quick Start

1. Get an API key

Top up at fetcher.sh/topup — a bby_live_… key is minted with the first credit purchase. There is no account or OAuth step.

2. Configure your MCP client

Claude Desktop

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

{
  "mcpServers": {
    "fetcher": {
      "command": "npx",
      "args": ["-y", "@fetcher-sh/mcp-server"],
      "env": {
        "FETCHER_API_KEY": "bby_live_..."
      }
    }
  }
}

Restart Claude Desktop. The catalog tools will be available in any chat — Claude will pick the right one based on your prompt.

Cursor

Open Settings → MCP → Add new MCP Server:

{
  "mcpServers": {
    "fetcher": {
      "command": "npx",
      "args": ["-y", "@fetcher-sh/mcp-server"],
      "env": {
        "FETCHER_API_KEY": "bby_live_..."
      }
    }
  }
}

VS Code (Copilot Chat with MCP)

Add to your MCP servers config — same shape as Claude Desktop / Cursor.

Claude Code

claude mcp add fetcher npx -- -y @fetcher-sh/mcp-server -e FETCHER_API_KEY=bby_live_...

3. Use it

In any MCP-enabled chat:

"Find recent tweets from @OpenAI about agents in the last week"

"Search YouTube for lofi streams uploaded this week"

"What are the top Reddit posts about stablecoin payments?"

The client will pick search_endpointsdescribe_endpointfetch_data (or a named service tool if FETCHER_SERVICE is set).

Authentication

Authentication is via the FETCHER_API_KEY environment variable, injected by your MCP client. The server never stores or logs the key. Each paid call sends the key as Authorization: Bearer to https://fetcher.sh.

search_endpoints and describe_endpoint work without a key — they read the bundled catalog.

Optional env vars:

Variable

Purpose

FETCHER_API_KEY

Prepaid bby_live_… key

FETCHER_SERVICE

Restrict to one service slug (twitter, youtube, …)

FETCHER_API_BASE

Override the API origin (default https://fetcher.sh)

Hosted Streamable HTTP (no install) is also available at https://mcp.fetcher.sh — that surface accepts the same Bearer key, or per-call USDC via x402.

Pagination

List endpoints (followers, replies, search results, comments, …) accept a cursor parameter and return a next_cursor (or cursor) field. Pass it back on the next call to page through.

Error handling

  • 429 / 5xx responses are automatically retried with exponential backoff (3 attempts, 1s / 2s / 4s)

  • Network timeouts: 30s per request

  • 4xx errors (other than 429) surface immediately to the model with the original message

  • Unknown paths and missing required parameters are rejected before any charge

Tools — full spec

Each tool's input schema is exposed via MCP's tools/list and follows JSON Schema. Browse interactively:

npx @modelcontextprotocol/inspector npx -y @fetcher-sh/mcp-server

What's NOT included

By design, this server exposes read-only public data. The following are intentionally excluded:

  • Posting, liking, following, DMs, or any write action on the underlying platforms

  • Official platform developer accounts / OAuth

  • Account login or 2FA

  • Per-call wallet payment (x402) on this stdio transport — use mcp.fetcher.sh or fetcher.sh/topup for that

Spec compliance

  • Built on @modelcontextprotocol/sdk v1

  • Targets MCP spec 2025-11-25

  • Transport: stdio (hosted Streamable HTTP lives at https://mcp.fetcher.sh)

  • Tested with: mcp-inspector, Claude Desktop, Cursor, Claude Code

Development

git clone <this-repo>
cd fetcher-mcp-server
npm install
npm run build
npm test
FETCHER_API_KEY=bby_live_... npm run inspect

License

MIT © fetcher.sh

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.

No tool schema history has been recorded yet.

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

  • A
    license
    Not graded
    quality
    D
    maintenance
    Pay-per-call web search for AI agents, settled in USDC on Base via the x402 protocol. No API key or subscription required; users fund a wallet and get a web_search tool.
    14
    1
    MIT
  • A
    license
    Not graded
    quality
    D
    maintenance
    Enables web search for AI agents with pay-per-search in USDC, no API keys needed.
    MIT
  • A
    license
    Not graded
    quality
    B
    maintenance
    Provides live data API for AI agents with 51 REST endpoints across finance, travel, weather, news, and more, using USDC micropayments without API keys.
    115
    MIT
  • A
    license
    Not graded
    quality
    C
    maintenance
    Enables AI agents to retrieve X/Twitter profiles and tweets, YouTube video and channel data, and TikTok profile and video stats on a pay-per-result basis without requiring login or platform API keys.
    0
    MIT

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/fetcher-sh/fetcher-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server