MCP Gateway
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
MCP Gateway
Recruiter hook: Tool-selection accuracy 61% → 94%. Same agent, 12× fewer tool tokens.
Production control plane for AI agents using MCP. Registers 15+ downstream servers (filesystem, GitHub, Postgres, fetch, …), semantically filters 200+ tools to top-8 per request, enforces RBAC, and logs every call to Postgres.
Problem
Your company has multiple MCP servers exposing 200+ tools. Dumping every schema into agent context means wrong tool picks half the time and thousands of wasted tokens on definitions never used.
Related MCP server: Peta Core
Solution
Agent → FastAPI Gateway → [Semantic Search → RBAC → Rate Limit → Postgres Audit]
↓ top-8 tools
┌────────────────────┼────────────────────┐
▼ ▼ ▼
filesystem MCP GitHub MCP Postgres MCP
fetch MCP GitLab (collision) + extension serversFeatures
# | Feature | Module |
1 | Real MCP servers |
|
2 | Gateway registration |
|
3 | Semantic tool search |
|
4 | RBAC |
|
5 | Audit log |
|
6 | Rate limiting | Token bucket per caller + tool |
7 | Benchmark | 50 queries — naive vs semantic vs role-scoped |
8 | RBAC benchmark | 30 unauthorized calls — 100% blocked |
Metrics
Tool-selection accuracy: 61% → 94% (semantic top-8)
Tool-definition tokens: 12× reduction
RBAC: 100% of unauthorized calls blocked
Quick Start
# With Docker (Postgres + gateway + MCP servers)
cp .env.example .env # add OPENAI_API_KEY, GITHUB_TOKEN (optional)
docker compose up -d
# Local dev
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
python -m gateway.mainSkills (reusable modules)
skills/mcp_client— connect to downstream MCP servers, health-checkskills/llm_adapter— OpenAI embeddings for semantic tool searchskills/audit_trail— Postgres audit store with sensitive arg redaction
Benchmarks
python benchmarks/run_benchmark.py # 50-query accuracy benchmark
python benchmarks/run_rbac_benchmark.py # 30 unauthorized callsDemo GIF
# Terminal 1: start gateway
python -m gateway.main
# Terminal 2: run demo flow (record with ScreenToGif / OBS)
./scripts/demo_flow.ps1Demo Users
Role | Password | |
admin | demo123 | |
engineer | demo123 | |
sre | demo123 | |
readonly | demo123 |
Stack
FastAPI · MCP SDK 2.0 · OpenAI embeddings · Postgres · Docker Compose · OAuth2/JWT
License
MIT
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
An MCP gateway for teams deploying AI tools at scale to employee-facing and headless agents.
MCP gateway: one connection for your AI apps, with tool policies, PII redaction and audit logs.
MCP gateway with runtime security policy, tool-call-level control, and audit of agent actions.
Zero-setup MCP gateway securely connecting AI to your tools with authentication and workflows
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceA universal gateway that aggregates multiple MCP servers into a single interface while providing advanced token optimization, result filtering, and automated summarization. It enables efficient management of large tool catalogs and reduces context usage by up to 95% for major AI clients.18MIT
- FlicenseNot gradedqualityBmaintenanceA production-ready MCP gateway and control plane that provides credential vault, policy engine, audit logging, and managed runtime for routing tool calls between AI agents and downstream MCP servers.59-
- AlicenseNot gradedqualityDmaintenanceA production-grade MCP server designed for multi-tenant, authenticated, and observable AI agent systems, enabling secure tool execution across heterogeneous data sources.68MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to discover and execute tools via a secure MCP server with JWT authentication, RBAC, rate limiting, and audit logging.1MIT