Filament/How it works

How it works.

Three steps from zero to a fully routed, observable AI stack.

Get your key

Free. No credit card. Click "Get API Key", enter your email, and your key is generated instantly in the browser. Store it as an environment variable.

Key format: fl- + 32 hex chars. Generated client-side.

bash

export FILAMENT_KEY="fl-a3f9c2d8e1b5f7a4..."

Route any model

Send one request to POST /api/route. Filament picks the model automatically or you pin one. If the primary provider fails, it falls back to the next one silently.

Auto-routing: code → GPT-4o, long reasoning → Claude, factual → Gemini.

bash

curl -X POST https://filament.works/api/route \
  -H "Authorization: Bearer $FILAMENT_KEY" \
  -d '{"model":"auto","prompt":"What is RAG?"}'

Observe everything

Every call is traced automatically. Pull logs via GET /api/observe or view them in the dashboard. Filter by session ID to trace a full conversation.

Logs include: model used, tokens, latency, status, error (if any).

bash

curl "https://filament.works/api/observe?limit=50" \
  -H "Authorization: Bearer $FILAMENT_KEY"

Ready to route your first prompt?

Get API Key →