Skip to main content

​
Option A — Zero‑Code via OpenAI Proxy

Point your batch/files/completions to our proxy and keep your existing OpenAI SDK code. 
# example: chat completions
curl -N -X POST "$BASE/api/providers/open_ai/v1/chat/completions"   -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json"   -d '{"model":"gpt-4o-mini","messages":[{"role":"user","content":"Hello"}],"stream":true}'
→ Full details: /http/openai-proxy

​
Option B — Auto‑Instrument (OTEL)

Emit spans from OpenAI, HTTP, and tooling.

​
Option C — Semantic Steps (Optional)

Use Traceloop/OpenLLMetry to add workflow/task/agent/tool spans that group low‑level spans into readable steps. If you skip this, Brixo still infers steps from raw spans.