Command center✦
Everything you need to run, route, and monitor your AI models.
Your models.
Your rules.
One intelligent gateway for every free tier and AI provider. Your data, your server, your control.
Request activity
Usage over the last 24 hours
Provider health
Real-time routing availability
One API. Every model.
Connect OpenClaw, Hermes, custom apps, and coding agents using the same OpenAI-compatible endpoint.
/v1/chat/completionsModel providers
Bring your own API keys. Secrets are encrypted at rest.
Qwen3.8 Max → Qwen3.8 27B → anonymous OVH fallback
Free doesn't mean unlimited
Atlas Relay respects each upstream provider's own quotas and terms. Keys remain on your server and are never returned to the browser.
Model explorer
Browse upstream models without a paid catalog. Discover what's available directly from each provider.
Live, without a catalog subscription
Model IDs are fetched from the provider's own API. Catalog entries may include models outside chat completions or paid models. Free eligibility is determined by your upstream account.
AI playground
Test live models, inspect routing, and measure response times.
Try a prompt
Built for Qwen3.8
Qwen3.8 Max is preferred when its xKiro API key is connected. Otherwise the router chooses the available 27B endpoint.
Anonymous OVH is available immediately, but reasoning-heavy responses can be slow.
API keys
Give each application its own revocable access token.
Send your key in the Authorization header. Never expose it in frontend code.
Your application keys
Keys are only shown once when created
Request activity
Live throughput, latency, and routing decisions across every provider.
Recent requests
Latest 35 attempts. Failed requests are retained for diagnosis.
| Time | Provider | Route | HTTP | Latency | Tokens |
|---|
Gateway settings
Adjust routing preferences and connect external agents.
Inference settings
These settings apply to your personal AI gateway.
Connect your agents
Configure any OpenAI-compatible application with these details:
qwen3.8Bearer your API keyAbout Atlas Relay
Independent self-hosted gateway. No paid catalog subscription required. Model availability is based on each provider's real API access.