Configure which models power each pipeline stage — cost, quality, and latency trade-offs
Control plane offline
Start the control plane to see model routing data.
Terminal: make up
or System Health → Start Service
Installed local models and configured cloud providers
| Provider | Model | Local | Size | $/1K in | $/1K out | Avg Latency | Calls |
|---|
Which model handles each task type — defaults and per-task overrides
| Task Type | Class | Provider | Model | Source | Est. Cost/Call |
|---|
To change routing: use the API PUT /intelligence/routing-policies or run scripts/seed_routing_policies.py. UI editing coming in next iteration.
Current spending based on model routing — all local models are free
No model calls recorded yet.
Compare output quality between models on the same task — scored on 5 dimensions