Briefcase AI
Govern every decision your AI makes. Enforce controls before an action runs, capture the full context, and keep a record you can verify later.
pip install briefcase-ai==4.4.0 What do you want to do?
Watch a decision live Instrument a function in about five minutes and stream records to memory or the console. Gate a model call Enforce entitlements, quota, cooldowns, and typed failure outcomes in Python or TypeScript. Protect an audit trail Chain records together, detect concurrent conflicts, and verify a history from a trusted anchor. Store and replay Save a DecisionSnapshot, reload it by ID, and compare a replay against the original output. Control an action Deny-by-default guardrails and versioned routing before the model acts. Prove what happened Reconstruct evidence and policy as-of a past moment, then seal a verifiable bundle.
How a decision moves
Briefcase AI organizes around one lifecycle. Use it as a learning path inside Capabilities, then use package-specific API reference for exact signatures.
- Capture
- Control & Route
- Store & Query
- Replay & Verify
- Prove
- Operate
Prefer a walkthrough? Start with Audit a Decision End-to-End or jump to the Python API reference.
Separate product
Looking for model, fine-tune, and dataset lineage? Artifact Graph (oci-bai) lives at oci.briefcaseai.io and is also filed under More → Artifact Graph in these docs.