Self-hosted MCP gateway for coding agents — one place for all your MCP servers, one endpoint per agent, per-agent tool switchboard with OAuth handled for you.
A small self-hosted switchboard for MCP: put all your MCP servers in one place and hand each of your coding agents (Claude Code, Codex, …) its own endpoint with exactly the servers you want it to see.
- One-click templates — add 20+ popular servers (GitHub, Linear, Jira & Confluence, Notion, Stripe, Shopify, Vercel, Netlify, Cloudflare, Supabase, AWS, Zapier, Gmail, Google Drive/Calendar/Sheets, Context7, shadcn/ui, Playwright, Chrome DevTools, Firecrawl, Exa, Brave Search) from a visual picker; OAuth servers go straight to the browser consent page, the rest ask only for the API key they need. - Local servers — stdio processes (npx …, uvx …) spawned and supervised by the switchboard, with restart-on-crash and stderr logs. - Remote servers — Streamable HTTP or SSE, with bearer-token, custom-header, or full OAuth 2.1 auth. OAuth tokens are stored encrypted and refreshed proactively in the background so auth never goes stale. - Per-agent switch matrix — enable/disable each server per agent; changes apply live via tools/listchanged, no agent restart. - Paste to import — paste a claude mcp add … command or an mcpServers JSON block into Add server → Paste config; the switchboard parses it (multiple entries supported) and shows a preview before creating. - Namespaced tools — githubcreateissue, google-workgmailsearch. Multiple accounts of the same service are just multiple server entries with different slugs. - Agents know which account is which — give each server a description ("Work Gmail — carl@company.com") and the switchboard weaves it into every tool description, the server
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/cmer--mcp-switchboard.html)
ChatGPT thinks. Codex works. Use ChatGPT as the planning brain while keeping the Codex harness.
shinthink/blitzstrike⚡ Blitz Strike — a universal MCP penetration-testing toolbelt. Structured methodology: reconnaissance & attack-surface mapping, source-to-sink analysis, and live validation. 57 escalation chains, 130-
0xwilliamortiz/openclaude-improvedruns anywhere. uses anything
halofyai/halofyHalofy is the open access and governance layer for AI agents across your organization. Identity, policy, provenance, audit, and signed erasure.
CaptureGrubEnchant/SolidWorksSolidWorks MCP Server connects an AI assistant to a running SolidWorks instance. Sketch, extrude, fillet, export STEP/STL, generate macros.
arielshad/3d-asset-server3d assets api and mcp
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.