RelayForge AI
Unified AI Gateway
FeaturesArchitectureDocsStatus
Get Started
RelayForge AI
Unified AI Gateway
Workspace
Public Sandbox
Cloudflare Pages + Workers
PlaygroundProvider StatusLogsUsageSettingsDocs
System mode
demo
Fallback chain active
System mode
demo
Free-tier routing with automatic fallback and demo-safe reliability.
Marketing Site
Core Experience

Run prompts through the live gateway workspace

This screen uses the real chat and streaming endpoints. The Worker returns the response while the interface surfaces strategy, fallback, final provider and latency in one place.

POST /api/v1/stream
streaming on

Prompt composer

One payload, multiple routing strategies and a real streaming transport.

auto
Mode
demo
Streaming
Enabled by default
Fallback order
1. Groq
2. SambaNova
3. Cerebras
4. Gemini
5. OpenRouter
6. Mock
Public demo
If upstream providers fail, mock keeps the public interface testable.

Streaming response

Live token output, metadata and errors without forcing a layout change.

Empty

No response yet

Run a request to inspect stream tokens, provider metadata and normalized fallback events.

Request metadata

Selected strategyauto
Started atIdle
Transporttext/event-stream

Response metadata

Attempted providerPending
Final providerPending
LatencyPending
FallbackNo

Provider rail

The current snapshot comes from the Worker status endpoint and verifies readiness across the provider chain.