# IBM/mcp-context-forge

An AI Gateway, registry, and proxy that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails and management. Optimizes Agent & Tool calling, and supports plugins.

Repository: https://github.com/IBM/mcp-context-forge
Canonical: https://ross.abutalabs.com/products/mcp-context-forge
Homepage: https://ibm.github.io/mcp-context-forge/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agents, ai, api-gateway, asyncio, authentication-middleware, devops, docker, fastapi, federation, gateway, generative-ai, jwt, kubernetes, llm-agents, mcp, model-context-protocol, observability, prompt-engineering, python, tools
Last push: 2026-08-26T18:27:58+00:00

## Health v2 (maintenance only)
Score: 86/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 98, longevity 34
- inputs: {"age_days": 482, "days_push": 7, "days_rel": 15, "gap_med": 14.0, "n_releases_24m": 25}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4372, forks 839 (observed 2026-08-28T04:08:46.766602+00:00)

## What it is
ContextForge is an open-source AI gateway, registry, and proxy from IBM that federates MCP, A2A, and REST/gRPC APIs behind a unified endpoint with centralized discovery, guardrails, auth, and observability. It optimizes agent and tool calling, supports 40+ plugins, and deploys via PyPI or Docker with Redis-backed federation on Kubernetes.

## Use cases
- expose all my MCP servers through one gateway endpoint
- proxy and federate REST and gRPC APIs for AI agents
- add rate limiting and authentication to MCP tool calls
- centralize discovery and management of agents and tools
- add OpenTelemetry tracing to LLM agent traffic
- run a multi-cluster MCP registry on Kubernetes
- translate gRPC or REST services into MCP tools
- apply guardrails and plugins to agent tool calling

## When to choose
- you need a unified gateway in front of many MCP, A2A, or REST/gRPC endpoints
- you want centralized auth, rate limiting, guardrails, and observability for AI tool traffic
- you deploy on Kubernetes and need federated, scalable tool registries
- you want plugin extensibility for custom transports and protocols

## When to avoid
- you only need a single simple MCP server with no gateway features
- you want a lightweight client-side SDK rather than a self-hosted service
- your stack is not Python/Docker friendly and you need a managed SaaS gateway

## Facets
- artifact type: service
- maturity: active
- function: api-gateway, proxy, mcp, agent-framework, middleware, auth, rate-limiting, tracing, monitoring, caching, plugin-system, webhook
- domain: artificial-intelligence, large-language-models, apis, developer-tools, self-hosted, backend
- platform: python, self-hosted, cloud
- tags: mcp-gateway, a2a-protocol, tool-registry, federation, guardrails, opentelemetry, fastapi, grpc, rest-proxy, ai-gateway, ai-agents, devops, docker, kubernetes, web-server

## Member repositories
- IBM/mcp-context-forge (main) score 86

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:46.766602+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:21:21.334933+00:00, confidence not recorded.
  - readme: https://github.com/IBM/mcp-context-forge (fetched 2026-08-28T04:08:46.766602+00:00, sha 11ee52c55019)
  - homepage: https://ibm.github.io/mcp-context-forge/ (fetched 2026-08-29T09:09:46.064513+00:00, sha 36c6c3c2e4f9)
- Data as of 2026-08-30T08:39:29.467469+00:00.