Application teams get one consistent, OpenAI-compatible API for every model and tool, from hosted providers to self-hosted inference.
Platform teams keep credentials, quotas, routing, failover, and usage — managed centrally, enforced by Envoy.
Agent Router configures. Envoy handles the traffic.
An Agentic AI Foundation project · Built on Envoy · Formerly Envoy AI Gateway
Get Started · Docs · Blog · Talks · Discord
Tip
A ⭐ on agent-router helps other teams find the project. Thank you!
Try it with one command on your laptop. Ship the same configuration on Kubernetes — no rewrite in between.
OPENAI_API_KEY=sk-... aigw runThen point anything OpenAI-compatible at http://localhost:1975/v1. → CLI guide · Kubernetes guide
| 💬 Discord | Questions, ideas, and the maintainers — join the server. |
| 🗓️ Weekly meeting |
Mondays · 8:00 am Pacific · 11:00 am Eastern · 4:00 pm London · 5:00 pm CET Open to everyone. Skipped on US public holidays. Agenda & notes · Zoom · Recordings |
| 🛠️ Contribute | Contributing guide · Good first issues · Discussions · Security policy · Maintainers |
| 📣 Follow | X · Substack · LinkedIn · Talks |
Using Agent Router? Add your logo to the adopters list. We follow the LF Projects Code of Conduct.
Agent Router is an Agentic AI Foundation project, powered by Envoy. Copyright © Agent Router a Series of LF Projects, LLC. See the trademark policy and lfprojects.org for terms of use and other project policies.
