Developers

One OpenAI-compatible endpoint

If your workspace's app already talks to OpenAI, it already talks to Retanu — change the base URL and key, nothing else. Routes across OpenAI, Anthropic, DeepInfra, and Google (Gemini).

Drop-in

cURL

# Host is illustrative — your workspace gets a branded subdomain
curl https://acme.api.retanu.com/v1/your-client/chat/completions \
  -H "Authorization: Bearer rtn_live_your-client_…" \
  -H "Content-Type: application/json" \
  -d '{"messages":[{"role":"user","content":"Summarize this ticket"}]}'

Python (openai SDK)

from openai import OpenAI

# Host is illustrative — your workspace gets a branded subdomain
client = OpenAI(
    base_url="https://acme.api.retanu.com/v1/your-client",
    api_key="rtn_live_your-client_…",
)

resp = client.chat.completions.create
    model="auto",                      # Retanu routes by task
    messages=[{"role": "user", "content": "…"}],
)

What you get back

A standard OpenAI-shaped completion, plus an x_retanu block with the routing decision — which provider and tier served the request, the cost, and the latency — so your telemetry sees exactly what happened.