What is PandaProbe?
PandaProbe is an open-source agent engineering platform that gives you deep observability into AI agent applications. Use it to trace, evaluate, monitor, and debug your AI agents in development and production — and, with the harness, let your agents heal themselves.Why PandaProbe?
PandaProbe is a unified platform for the full agent development lifecycle — from first run to continuous improvement. PandaProbe provides:Tracing
Capture LLM calls, MCP calls, tool usage, workflow steps, and custom agent logic as structured traces and spans.
Evaluation
Score traces and sessions to measure quality, reliability, regressions, and agent behavior over time.
Monitoring
Schedule recurring evaluations so new traces and sessions are checked automatically in production.
Self-Healing Harness
Wrap your agent in an operational envelope that detects quality degradations, lets the agent diagnose its own failures, and validates the operating rules it learns — automatically, with no human in the loop.
Three layers of tracing
- LLM Providers — zero-code wrappers for major providers
- Agent Frameworks — automatic integrations for agent frameworks
- Manual instrumentation — full control when you need it
Architecture
See how tracing, evaluation, monitoring, and the harness fit together
Quickstart
Trace your first LLM call in under 2 minutes
Tracing Overview
Understand the three layers of tracing
Installation
Install the SDK with optional extras
Harness
Turn any traced agent into a self-healing agent
Harness Quickstart
Wire the self-healing loop into your agent in under 5 minutes