a3m-router vs llm-d-router
LLM request routing projects compared
a3m-router is a self-hostable gateway for routing OpenAI-compatible requests across model providers. llm-d-router routes inference traffic within Kubernetes serving setups, using live serving signals and proxy integrations; the projects differ mainly in deployment context and routing goals.
a3m-router: Route LLM Requests Across Models
A3M Router is a TypeScript, self-hosted LLM gateway built around Jev System One single-pass routing. It directs OpenAI-compatible requests to models and providers based on the task, with failover and optional parallel ensembles.
llm-d-router: Route LLM Inference with Envoy and Gateway API
llm-d-router is a Go-based router for LLM inference traffic, using an Endpoint Picker with Envoy and the Kubernetes Gateway API. It routes requests using signals such as model-server load and KV-cache locality, with options for standalone or gateway deployments.
| a3m-router | llm-d-router | |
|---|---|---|
| Language | TypeScript | Go |
| License | MIT | Apache-2.0 |
| Stars | 17 | 422 |
| Forks | 5 | 472 |
| Last analyzed | Oct 23, 2026 | Oct 20, 2026 |
Key differences
- a3m-router is written in TypeScript and focuses on selecting among external model providers; llm-d-router is written in Go and selects inference endpoints using serving state.
- a3m-router offers heuristic and ML-based routing, provider fallback, semantic caching, and optional parallel ensembles; llm-d-router scores endpoints using KV-cache locality, load, and request priority.
- a3m-router exposes an OpenAI-compatible API for application integrations; llm-d-router integrates with L7 proxies such as Envoy through ext-proc and can connect to Gateway API infrastructure.
- a3m-router is intended for teams operating a self-hosted gateway; llm-d-router targets teams running LLM inference workloads in Kubernetes with a supported proxy setup.
- a3m-router is MIT-licensed and llm-d-router is Apache-2.0-licensed.
- The supplied repository figures list 16 stars and 5 forks for a3m-router, and 373 stars and 410 forks for llm-d-router. Both are listed as not archived.
Choose a3m-router if you…
- need a self-hostable OpenAI-compatible gateway across model providers.
- want provider fallback, semantic caching, or optional parallel ensembles.
- want heuristic or ML-based routing to help balance capability, cost, and availability.
Choose llm-d-router if you…
- run LLM serving workloads in Kubernetes and need routing informed by cache locality, load, or priority.
- use Envoy or Gateway API infrastructure to connect inference endpoints.
- need inference objectives, model-name rewrites for canary or A/B traffic, or coordination for multi-stage inference workflows.
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.