inference-gateway vs router
Go proxies for routing requests across AI providers
inference-gateway and router are Go proxies that give clients a shared API for requests to multiple AI providers. inference-gateway emphasizes provider compatibility, MCP tools, authentication, guardrails, and observability; router emphasizes scored model selection, routing inspection, and a self-hosted dashboard.
inference-gateway: Unify LLM Providers Behind One API
A Go-based, cloud-native gateway that puts multiple LLM providers behind one OpenAI-compatible API. It supports cloud and self-hosted models, with integrations for MCP, A2A, guardrails and OpenTelemetry.
router: Route Agent Prompts with Avengers-Pro Scoring
Weave Router is a Go proxy for Claude Code, Codex, Cursor, and other clients, routing each request across model providers. Its in-process cluster scorer, derived from the Avengers-Pro research, uses an on-box embedder to choose a model per action.
| inference-gateway | router | |
|---|---|---|
| Language | Go | Go |
| License | Apache-2.0 | Apache-2.0 |
| Stars | 215 | 5.6k |
| Forks | 48 | 159 |
| Last analyzed | Oct 23, 2026 | Oct 20, 2026 |
Key differences
- inference-gateway supports OpenAI-compatible chat across cloud and local providers, plus Anthropic Messages and OpenAI Responses endpoints; router accepts Anthropic Messages, OpenAI Chat Completions, and Gemini formats.
- inference-gateway can connect MCP tool servers and provides OIDC authentication, authorization, TLS, and configurable guardrails; router offers an in-process cluster scorer and a route-inspection endpoint.
- inference-gateway lists OpenTelemetry metrics, Prometheus export, and OTLP metrics push; router exports traces through OTLP and provides routing analytics endpoints.
- inference-gateway offers container and Kubernetes deployment options, including a companion operator; router offers a self-hosted dashboard and stores provider keys encrypted at rest when self-hosted.
- Both projects use Go and list Apache-2.0, though router notes that the license applies to releases beginning with router-v0.2.24, while earlier releases retain their shipped licenses.
- The listed project metrics differ: inference-gateway has 214 stars and 50 forks, while router has 5.6k stars and 151 forks.
Choose inference-gateway if you…
- need MCP tool-server integration or configurable authentication and request guardrails.
- want a common API for cloud and local providers, with container or Kubernetes deployment options.
- need metrics via Prometheus or OpenTelemetry for a self-hosted gateway.
Choose router if you…
- want requests routed to enabled models using an in-process cluster scorer.
- need to inspect routing decisions without calling an upstream provider.
- want a self-hosted dashboard and provider keys encrypted at rest on your own infrastructure.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.