Invariance is an AI-native observability platform designed to monitor AI agent systems and traditional software stacks. It automatically learns baselines, detects anomalies, correlates failures to deploys and code changes, and creates findings with full evidence. Key features include:
- Soma: AI-native observability that watches every agent run, catches failures, and provides root cause analysis with cited evidence.
- GPS: Open-source repo memory for coding agents, enabling context-aware code fixes.
- Automatic monitoring: No thresholds to tune; Soma learns normal behavior and alerts on drift.
- Findings: Detailed reports linking errors to specific deploys, code lines, and owners.
- Fix handoff: Drafts guarded changes ready for coding agents.
- Knowledge graph: Answers questions by grounding in Slack, GitHub, Sentry, and more.
Use cases include AI agent observability, LLM tracing, workflow monitoring, and ops engineering management.
