Observability

See everything your AI agents do

Traces, conversations, metrics, and a real-time monitoring dashboard. Know exactly what your agents are doing, how they perform, and what they cost.

Live

Available today

Ship agents to production with confidence. These features are live and ready to use.

Execution Traces

Every tool call, LLM request, and agent step captured with full context. Filter by status, model, agent, duration, and tokens.

Conversations

Group related traces into conversations. See the full interaction flow, trace count, and duration at a glance.

Monitoring Dashboard

Real-time timeseries charts, top spans analysis, and status distributions. Spot regressions before they hit users.

Agent Performance

Per-agent success rates, average and P95 latency, token usage, tool call counts, and error rates.

Cost Tracking

Track LLM spend per trace, agent, and model. Identify which agents drive cost and optimize accordingly.

Environment Filtering

Separate production, staging, and development data. Filter everything by environment with a single click.

Coming Soon

On the roadmap

These features are in active development. Design partners get early access.

Failure Clustering

Automatically group failures by root cause. See patterns across agents instead of investigating one-by-one.

Checkpoint Replay

Rewind to any failed step and replay with different inputs. Debug without re-running the full agent.

Semantic Search

Search traces by meaning, not just keywords. Find every trace where an agent discussed a specific topic.

Edit Replay

Step through an agent's execution frame by frame. See exactly what happened and when.

Start seeing what your agents do

Add two decorators. Get full observability in minutes. Free tier available — no credit card required.