glp1forum-mcp
MCP server that searches and reads https://glp1forum.com (a XenForo forum) via system curl, providing tools for full-text search, thread reading, image downloads, and forum browsing without any API keys.
README
glp1forum-mcp
An MCP stdio server that searches and reads https://glp1forum.com (a XenForo forum) with no API keys. It shells out to system curl because Cloudflare there blocks Node's own TLS fingerprint but passes a real curl with a browser User-Agent.
⚠️ The one gotcha
Verified only with Windows curl 8.18 (Schannel). Linux/macOS curl presents a different TLS fingerprint and may get a 403 from Cloudflare. If every request 403s, this is almost certainly why.
Requirements
- Node 18+
- System
curlon yourPATH(curl --versionshould work)
Tools
search_forum— full-text search with XenForo filters (keywords, nodes, author, date range, min replies, order, page,maxPagesup to 3). Keywords are AND-matched; category nodes needincludeChildNodes: true.get_thread— read a thread's posts (author, datetime, body, permalink), paginated.get_thread_images— opt-in: download a thread's image attachments at full size as image blocks so vision can read image-only pricing tables / stock boards / COA figures.maxcaps the count (default 4, max 5 — full-size images are token-expensive).list_threads— browse a forum section's thread list bynodeId, paginated.list_forums— list forum sections with their numeric node IDs (feed these tosearch_forum.nodesandlist_threads.nodeId).
Rate limiting
All requests share a global throttle: ≥2.5s between hits. On HTTP 403/429/503 the server backs off 8s plus up to 10s of jitter and retries once; if still blocked it throws glp1forum rate-limited or unavailable (HTTP 503) — wait ~60s before retrying.
The backoff is deliberately much shorter than the ~60s a Cloudflare 503 actually takes to clear: a handler that slept that long would trip the MCP client's own 60s timeout (-32001), and the sleep holds the search mutex. So it fails fast with a retryable error and asks the caller to wait instead — hence "wait ~60s" in the message.
Two env vars tune this:
GLP1_BACKOFF_MS(default8000) — the retry backoff. Raise it if one retry isn't enough.GLP1_INTERVAL_MS(default2500) — the global gap between requests. Raise it for heavy concurrent use; each search costs 2 requests, and 8-way concurrency trips Cloudflare at 2.5s.
Install
Option A — Claude Code plugin marketplace (recommended)
/plugin marketplace add Sedation6612/glp1forum-mcp
Then install the glp1forum-mcp plugin. This auto-wires the MCP server; the plugin's config uses ${CLAUDE_PLUGIN_ROOT} so there are no paths to edit.
Option B — Desktop Extension (.mcpb, one-click Claude Desktop)
Download glp1forum-mcp.mcpb from the repo's Releases and double-click it — Claude Desktop
installs it with no paths to edit and no npm install (the server is bundled into a single
dependency-free dist/index.mjs). You still need system curl on your PATH.
Option C — manual (Claude Desktop)
Run npm install && npm run build first, then add to claude_desktop_config.json:
"glp1forum": { "command": "node", "args": ["C:\\path\\to\\glp1forum-mcp\\dist\\index.mjs"] }
MSIX Claude Desktop installs put this config under
%LOCALAPPDATA%\Packages\Claude_pzs8sxrjxfjjc\LocalCache\Roaming\Claude\, not plain%APPDATA%.
Build (contributors)
The distributed artifact is a single bundled dist/index.mjs (esbuild). After editing anything
in src/, rebuild and commit dist/:
npm install
npm run build # -> dist/index.mjs (what the plugin & .mcpb run)
npm run pack # build + produce glp1forum-mcp.mcpb for Claude Desktop
Self-test
npm run selftest # or: node src/selftest.js
Makes a few dozen seconds of live requests under the throttle. Failure means curl is being blocked or the site's HTML drifted.
License
All rights reserved (see LICENSE). The source is public for viewing and for contributions back to this repository via pull request. It is not open source — you may not redistribute it or publish your own version. See the LICENSE file for details.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
E2B
Using MCP to run code via e2b.