Every AI. One API.
A single OpenAI-compatible endpoint that fans out to Claude, GPT, Gemini, Veo, Sora, Suno — and 100+ more. Production-grade routing, transparent Credits pricing, zero vendor lock-in.
# 1. Drop-in OpenAI base URL curl https://orux.top/api/v1/chat/completions \ -H "Authorization: Bearer sk-app-…" \ -H "Content-Type: application/json" \ -d '{ "model": "claude-opus-4.7", "messages": [{ "role":"user","content":"Hello, world." }] }'
- 1{
- 2 "id": "chatcmpl-9aA…",
- 3 "model": "claude-opus-4.7",
- 4 "choices": [{ "message": { "content": "Hello, world." } }],
- 5 "usage": { "total_credits": 3.6, "latency_ms": 240 },
- 6 "x_orux": { "route": "auto", "fallback": "none" }
- 7}
One request. Every upstream.
Orux AI sits between your app and every model provider. Pick a model id — Orux AI resolves it to the cheapest healthy upstream, retries on failure, and bills in a single unit (Credits).
- OpenAI-compatible protocol · drop in as a base URL
- Sub-200ms median routing overhead
- Multi-channel failover — redundant upstreams behind every model
- HMAC-signed webhooks for async tasks (video / music / Midjourney)
Built for production.
Everything you need to ship AI features fast — without re-architecting every quarter.
Drop-in OpenAI
Same protocol. Same SDKs. Swap base URL and ship.
100+ models
Claude Opus 4.7 · GPT-5.5 · Gemini 3 Pro · Veo · Sora · Suno · Midjourney.
Multi-channel failover
Routes around upstream outages. One provider dark ≠ you down.
Transparent Credits
Fixed 1 USD = 200 Cr. No expiry. No monthly minimum.
Streaming + tools
SSE, function calling, JSON mode — 1:1 with OpenAI.
Real observability
Per-model usage, cost, error rate, p95 latency — out of the box.
Operate it like infrastructure.
Dashboard, docs and playground — engineered for the same humans who read RFCs.
Real-time dashboard
Live cost, calls, latency and error rate — per key, per model.
- Credits balance + burn rate
- 24h call & p95 latency charts
- Per-key credit quota + per-user concurrency
| time | model | cr | ms | status |
|---|---|---|---|---|
| 12:04 | claude-opus-4.7 | 3.6 | 240 | 200 |
| 12:03 | gpt-5.5 | 2.1 | 180 | 200 |
| 12:02 | gemini-3-pro | 1.8 | 210 | 200 |
| 12:01 | veo-3.1-quality | 180.0 | async | queued |
| 11:59 | deepseek-v3.1 | 0.4 | 160 | 200 |
Docs that respect your time
Endpoint reference, code samples in 4 languages, and a real OpenAPI spec you can fork.
- cURL · Node · Python · Go side-by-side
- Live request inspector
- OpenAPI 3.1 — downloadable
/api/v1/chat/completions
curl https://orux.top/api/v1/chat/completions \
-H "Authorization: Bearer sk-app-…" \
-H "Content-Type: application/json" \
-d '{ "model": "claude-opus-4.7",
"messages": [{"role":"user","content":"hi"}] }'Playground on your own key
Try any model with your real sk — what you see is exactly what your app gets.
- Streaming responses with first-byte timing
- Image, video and music generation
- Templates: 50+ curated prompts
// 200 req/min, burstable to 50 const limiter = tokenBucket(200, 50); await limiter.acquire(); return fetch(url, init);
Pay only for what you use.
Top up Credits when you need them. No subscription, no expiration, no surprise invoices.
| tier | credits | cost (USD) | note |
|---|---|---|---|
Free evaluation | 200 Cr | $0 | on signup · no card |
Starter side-project | 2,000 Cr | $10 | one-time top-up |
Growth most popular | 20,000 Cr | $100 | covers ~5k chat calls |
Scale production | 200,000 Cr | $1,000 | priority routing |
Ship faster. Outgrow vendor lock-in.
Email sign-up, no card. From signup to first response in under five minutes.