Key features
Core capabilities each platform advertises.
- Agent tracing with a readable transcript view
- Signals for automatic failure detection and clustering
- Evals and datasets built from production traces
- Full SQL access through the editor, CLI, and MCP
- Nested traces for every LLM call, tool run, and agent turn
- Threads for multi-turn sessions
- Online evals on live production traffic
- LLM-judge, code, and human evaluators
- Datasets from production logs and experiments at scale
- Monitors with Slack, email, and webhook alerts
- Prompt versioning and deployment without code changes