MCP MCP server · io.github.ENTERPILOT
io.github.ENTERPILOT/gomodel
Self-hosted gateway aggregating upstream MCP servers behind one authenticated HTTP endpoint.
- host stars
- 1,166
- 30-day movement
- +3211/day
- Related entries
- 1
- Connections
- 0
ocilocalstreamable-http
GoModel is a self-hosted gateway that aggregates multiple upstream MCP servers behind a single authenticated HTTP endpoint using streamable-http transport. It also accepts requests in two compatible formats and lists support for many LLM providers, including OpenAI, Anthropic, Google Gemini, and others.
Use it when you want one authenticated entry point in front of several MCP servers and LLM providers instead of wiring each client to each backend individually.
Use it to
- Consolidate upstream MCP servers behind one endpoint
- Add authentication in front of MCP backends
- Cache repeated prompts with exact or semantic caching
- Track token costs across providers
- Deploy infrastructure via the provided Docker Compose setup
For Teams self-hosting MCP infrastructure for multiple clients or providers
- Runs
- installed — on your machine
- Transport
- streamable-http
- Packaged as
- oci:docker.io/enterpilot/gomodel:0.1.93
- Install
- docker run docker.io/enterpilot/gomodel:0.1.93
- Needs
- 2 environment variables, including secrets
- Version
- 0.1.93
- Last release
- 2026-09-16
- Docker pulls, lifetime
- 50,683
topicsmcpgatewayself-hostedllm-providerscachingproxy