MCP Context Forge

Active
GitHub Python Apache-2.0

Description

An AI Gateway, registry, and proxy by IBM that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails, and management.

Key Features

  • Unified gateway federating MCP, A2A, and REST/gRPC APIs with a single endpoint and centralized governance
  • Tools Gateway with MCP, REST, gRPC-to-MCP translation, and TOON compression support
  • Agent Gateway supporting A2A protocol, OpenAI-compatible and Anthropic agent routing
  • 40+ plugins for additional transports, protocols, and integrations with extensible architecture
  • OpenTelemetry observability with Phoenix, Jaeger, Zipkin, and other OTLP backends for distributed tracing

Use Cases

💡 Centralized API gateway for enterprise AI infrastructure managing multiple MCP and REST services
💡 Virtualizing legacy gRPC or REST APIs as MCP-compliant tools without rewriting backend code
💡 Observability and rate-limiting across federated AI agent and tool calling pipelines
💡 Multi-cluster Kubernetes deployments with Redis-backed federation and caching for scaling

Tags

mcp a2a api-gateway registry guardrails ibm python

Categories

🔌 MCP Tools 🛡️ Security & Guardrails

Quick Start

uvx --from mcp-contextforge-gateway mcp-contextforge-gateway. Set BASIC_AUTH_PASSWORD=pass, MCPGATEWAY_UI_ENABLED=true, PLATFORM_ADMIN_EMAIL=admin@example.com, PLATFORM_ADMIN_PASSWORD=changeme. Or Docker: docker pull ghcr.io/ibm/mcp-context-forge. Admin UI at http://localhost:8080.
Visit GitHub Visit Website View Docs

Related Projects