2026-10-07

OpenAI-compatible endpoints that do not require an API key

Four chat listeners on this host accept the OpenAI body with no bearer token.

Several processes on this host speak the OpenAI chat body and none of them check a key. Port 8000 and port 8080 are the general listeners. Port 4000 is the LiteLLM-style proxy. Those three list gpt-4o-mini, claude-opus-5-5, claude-fable-5-1, and gpt-6-astra. Port 8642 is the Hermes-style API, model hermes-agent. Port 18789 is OpenClaw, model openclaw/default.

The path is always /v1/chat/completions for chat and /v1/models for the list. Omit the Authorization header. A client library that requires the field to be non-empty can send any placeholder. It is not checked.

Pick one base URL and stay there for a test. Switching ports mid-conversation only matters if you care which model name the process reports by default.

On this host

The same body works on each of those ports. This copy uses port 8000. No API key.

curl http://datattention.com:8000/v1/chat/completions \
  -H 'Content-Type: application/json' \
  -d '{"model":"gpt-4o-mini","messages":[{"role":"user","content":"Hello"}]}'

The full list of ports and paths is on the endpoints page.