Overview
ZizkaDB is an audit trail database for AI agents that provides tamper-evident, checksum-backed decision logs with session replay and time-travel debugging. It is built to support EU AI Act Article 12 record-keeping. The system links every agent step to the step that caused it, allowing users to trace the causal chain behind any event. Key features include causal lineage tracking, time-travel to past agent states, drift detection, and semantic search over agent history. It can be self-hosted on your own Postgres with one Docker command or used via ZizkaDB Cloud. The software is AGPL-3.0 licensed, with no per-trace billing for self-hosted deployments, and self-hosted data never leaves your infrastructure. It integrates with Python, TypeScript, LangChain, CrewAI, LiveKit, and MCP, and provides SDKs for logging events and querying causal chains.
Pricing
Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.
Visit WebsitePrices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.
Use Cases
- Teams use ZizkaDB to trace the causal chain behind any agent event with a single call.
- Teams use ZizkaDB to rebuild exactly what an agent knew at any past moment for time-travel debugging.
- Teams use ZizkaDB to maintain a tamper-evident audit trail for EU AI Act Article 12 compliance.
- Teams use ZizkaDB to detect when an agent's behavior shifts from its baseline.
- Teams use ZizkaDB to perform semantic search over an agent's history.
- Teams use ZizkaDB to inject relevant past events into the next prompt.
- Teams use ZizkaDB to comply with GDPR erasure requests by metadata filter.
Features
Record checksum-backed agent decision logs
Replay past agent sessions
Inspect historical agent knowledge
Trace causal chains behind events
Search agent history semantically
Detect behavior drift across sessions
Erase records by metadata filter
Reviews
No reviews yet. Be the first to share your experience!