mcp context forge

Report Abuse

Basic Information

ContextForge MCP Gateway is an open source gateway, registry and proxy for the Model Context Protocol (MCP) that provides a unified endpoint in front of MCP servers and REST APIs. It is designed to federate multiple MCP and REST services, virtualize legacy APIs as MCP-compliant tools, and expose tools, prompts and resources to AI clients through a single managed interface. The project is packaged for PyPI and OCI containers and can be run locally, in containers or scaled on Kubernetes with Redis-backed federation and caching. The README highlights deployment options, a stdio wrapper for clients, an optional Admin UI, OpenTelemetry observability and extensive configuration via environment variables. The repository is provided without official IBM support and the README cautions the current release is alpha and intended for development and experimentation rather than production deployment.

Links

Categorization

App Details

Features
The project includes multi-gateway federation and discovery with health checks and Redis syncing, virtualization of REST/gRPC services into virtual MCP servers, and a REST-to-MCP adapter that supports automatic JSON Schema extraction, header and token handling, retries, timeouts and rate-limit policies. It supports multiple transports including HTTP, JSON-RPC, WebSocket, Server-Sent Events, stdio and streamable HTTP. Management capabilities include CRUD APIs for tools, prompts, resources, gateways, servers and roots plus an Admin UI built with HTMX and Alpine.js. Built-in auth supports Basic and JWT schemes and encrypted tool auth values. Observability is vendor-agnostic OpenTelemetry with backends such as Phoenix, Jaeger and Zipkin. Tooling includes a stdio wrapper, dev container, Makefile targets, tests and doctest coverage.
Use Cases
For developers and operators building or integrating agentic systems, the gateway centralizes discovery, authentication, rate limiting and observability so clients can call many heterogeneous backends via one endpoint. It simplifies migrating or adapting legacy REST or gRPC services into MCP tools, enables creating virtual servers that bundle tools, and offers resource and prompt registries with Jinja2 templating and caching. The stdio wrapper lets MCP clients that cannot open authenticated SSE streams still access tools. OpenTelemetry instrumentation provides distributed tracing and LLM-specific metrics for debugging and cost analysis. The container images, PyPI package, Makefile, CI configuration and developer docs support local development, testing and deployment automation while the README warns users to perform their own security review for production use.

Please fill the required fields*