See what your agents
are actually doing
3 lines of code. Full traces, costs, and errors — visible in seconds. Stop flying blind.
Three steps to full observability
5 minutes to your first trace.
01
Install
Add the SDK to your project
npm install @trefur/observe # or pip install trefur-observe # or go get github.com/trefur/observe-go
02
Instrument
One call. Your agent is traced automatically.
from trefur_observe import TrefurObserve trefur = TrefurObserve(api_key="trf_live_...") trefur.instrument()
03
See Everything
Every run traced with full context
# Visit app.trefur.com to see traces
Everything you need to debug and monitor agents
Full Traces
See every decision your agent makes — inputs, outputs, timing, and context.
Cost Tracking
Know what you're spending per agent, per request, per model.
Error Debugging
Full stack traces and context when things break.
Latency Monitoring
Find the bottleneck in your agent chain. Optimize what's slow.
Alerts
Get notified when agent performance degrades or errors spike.
Works with Your Stack
Python, JavaScript, and Go SDKs. Get started in minutes regardless of your setup.
Instrument your agent in minutes
Initialize the SDK and your agent is traced. No manual spans. No complex setup.
- Automatic trace capture
- Cost per request
- Error context
- Latency breakdown
terminal