Overview
hermes proxy runs a local HTTP server that forwards OpenAI-compatible requests to an OAuth-authenticated provider. External apps can point at the proxy with any bearer token; the proxy attaches your real Hermes credentials, so any OpenAI-compatible client can use Hermes as a drop-in backend.
hermes proxy <start|status|providers>
Subcommands
start — Run the proxy in the foreground
hermes proxy start
hermes proxy start --provider nous --host 127.0.0.1 --port 8645
hermes proxy start --provider xai --port 9090
| Flag | Description |
|---|---|
--provider PROVIDER |
Upstream provider: nous or xai (default: nous). See hermes proxy providers. |
--host HOST |
Bind address (default: 127.0.0.1). Use 0.0.0.0 to expose on LAN. |
--port PORT |
Bind port (default: 8645). |
status — Show which proxy upstreams are ready
hermes proxy status
Prints the readiness state of configured upstream providers so you know whether the proxy can serve requests before you point a client at it.
providers — List available proxy upstream providers
hermes proxy providers
Shows the providers the proxy can use (currently nous and xai).
Common Scenarios
Use Hermes from another OpenAI-compatible tool
hermes proxy start --port 8080
# In the other tool:
# API base: http://localhost:8080/v1
# API key: any-value (the proxy handles auth)
Expose the proxy on your LAN
hermes proxy start --host 0.0.0.0 --port 8645
Switch upstream provider
hermes proxy providers
hermes proxy start --provider xai
Test the proxy with curl
hermes proxy start --port 8645 &
curl http://localhost:8645/v1/models
Tips
- The proxy is local by default (
127.0.0.1) — it does not expose anything to the internet unless you change--host. - Any bearer token value works in the client; the proxy attaches your real Hermes OAuth credentials.
- The proxy must run in the foreground; start it in a dedicated terminal or under a process manager for long-lived use.
- Use
hermes proxy statusto verify upstream readiness before connecting a client.