OPENAI-COMPATIBLE GATEWAY

Build with good models.
Skip the busywork.

One clean API for a small, curated model loadout—with streaming, safety screening, automatic failover, and honest limits.

API BASEloading…
START BUILDING VIEW JSON
OpenAI formatDrop-in chat and Responses routes
Failover built inauto tries healthy models in order
Privacy-mindedNo prompts stored in public metrics

AVAILABLE LOADOUT

Pick a model

Use auto for the recommended resilient route, or pin a model when you need a specific capability.

Loading models…

QUICKSTART

Your first request

cURL 
curl API_BASE_URL/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "messages": [{"role":"user","content":"Hello!"}],
    "stream": true
  }'

Anonymous requests need no key. Signed-in accounts use Authorization: Bearer andy_…. The X-Andy-Model response header identifies the model that actually answered.

LIVE SIGNAL

72-hour activity

VALID REQUESTS—
UPSTREAM SUCCESSES—
TOP MODEL—

Loading aggregate usage…

Checking your session…