One API across providers.
Integrate once with a consistent interface for chat, responses, embeddings, images, files, and more.
Propel Router
One secure gateway. Every model.
Route, govern, and observe every AI request through one production-ready API.
Gateway status
Healthy
All systems operational
Provider routing
Unified
Automatic failover ready
POST /v1/chat/completions
200 OK
curl https://api.example.com/v1/chat/completions \
-H "Authorization: Bearer $PROPEL_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model":
"best-available",
"messages": [
{
"role": "user",
"content": "Build something great."
}
]
}'
Built for production
Everything between your product and every provider.
A focused control layer for access, routing, cost, and reliability without changing your application.
Routing that keeps moving.
Balance requests across deployments with retries, fallbacks, health checks, and distributed coordination.
Every token, visible.
Understand usage and spend by key, user, team, model, and provider with persistent operational records.
Secure by default.
Centralize authentication, virtual keys, budgets, rate limits, and access policies at the gateway.
Clear architecture
Your application stays simple.
Propel Router handles the operational complexity while your product speaks one familiar API.
Your product
One integration
Propel Router
Auth · Route · Observe
AI providers
Every model
One response
Consistent format
Designed to stay up
Reliable when it matters most.
Health-gated releases, durable state, distributed coordination, streaming, and deliberate scaling form the foundation of every deployment.
PostgreSQL persistence
Redis coordination
Durable object storage
Streaming responses
WebSocket support
Health-gated deploys
One gateway.
Infinite possibilities.
Start building with a simpler, safer AI foundation.