# Briefcase AI > Govern AI decisions with controls, capture, replay, tamper evidence, and reproducible data across Python and TypeScript. Start with the Quickstart, then choose a task. Briefcase AI 4.4 ships Python plus focused TypeScript packages. A full-text version of every page is at https://docs.briefcaseai.org/llms-full.txt ## Start - [Overview](https://docs.briefcaseai.org/): Govern every decision your AI makes — enforce controls before an action runs, capture the full context behind it, and keep a complete record you can verify later. An open-source Python SDK. - [Quickstart](https://docs.briefcaseai.org/getting-started/quickstart/): Record a live decision with @capture and observe() in about five minutes using the base package. - [Installation](https://docs.briefcaseai.org/getting-started/installation/): Install Briefcase AI 4.4 for Python or TypeScript, then add only the optional providers your application uses. - [Core Concepts](https://docs.briefcaseai.org/getting-started/core-concepts/): The mental model behind Briefcase AI — why a decision, its inputs and outputs, its environment, and its grouping are recorded as distinct, reproducible records. - [Why Briefcase AI](https://docs.briefcaseai.org/getting-started/why-briefcase/): Briefcase AI is infrastructure for governing the decisions your AI systems make — enforce controls before an action runs, capture the full context behind every decision, and keep a complete, reproducible record you can verify later. - [Persist & Replay](https://docs.briefcaseai.org/getting-started/persist-and-replay/): Save a DecisionSnapshot to storage, reload it by ID, and compare a deterministic replay against the original output. - [AI-Assisted Setup](https://docs.briefcaseai.org/getting-started/ai-assisted-setup/): The fastest way to add Briefcase AI to a project — let your AI coding assistant do it. Point it at the docs, or give it Briefcase AI's tools over MCP. - [What's New in 4.4](https://docs.briefcaseai.org/getting-started/whats-new/): Briefcase AI 4.4 adds concurrent hash-chain safety on top of 4.3 integrity and the 4.2 controls, npm, instrumentation, and adapter expansion. ## Guides - [Audit a Decision](https://docs.briefcaseai.org/guides/audit-a-decision/): Follow one support-triage decision from the evidence it depended on, through the policy that governed it, to a sealed, tamper-evident record you can verify months later. - [Govern Agent Actions](https://docs.briefcaseai.org/guides/govern-agent-actions/): Put controls in front of an agent so an action is evaluated before it runs — deny-by-default guardrails, fail-closed pipelines, and a versioned policy that decides where the work goes. - [Observe in Production](https://docs.briefcaseai.org/guides/observe-in-production/): Wire up live observability for a running triage agent — emit every decision, watch token cost against a budget, measure output drift, trace with OpenTelemetry, and fire events when something looks off. - [Build Reproducible RAG](https://docs.briefcaseai.org/guides/reproducible-rag/): Make a retrieval-augmented decision reproducible — version the embedding index, snapshot the external data it read, and validate that a prompt's references resolve before a model ever sees them. - [Compare Model Versions](https://docs.briefcaseai.org/guides/run-an-evaluation/): Push a baseline and candidate through the gateway, compare them at every depth, and read the verdict scorecard. ## Capabilities - [Capture / Decision Recording](https://docs.briefcaseai.org/features/decision-recording/): Capture every AI decision — inputs, outputs, model, and a content fingerprint — so it can be replayed, audited, and proven later. - [Capture / Exporters](https://docs.briefcaseai.org/features/exporters/): Stream captured decisions to the console, a file, memory, or your own external sink as they happen. - [Capture / PII Sanitization](https://docs.briefcaseai.org/features/pii-sanitization/): Minimize sensitive data — detect and redact it before a decision record is ever stored. - [Control & Route / Invocation Controls](https://docs.briefcaseai.org/features/invocation-controls/): Enforce tenant caps, quotas, cooldowns, retries, and provider fallback before an AI call spends tokens. - [Control & Route / Guardrails](https://docs.briefcaseai.org/features/guardrails/): Enforce deny-by-default, fail-closed controls before an agent action runs. - [Control & Route / Routing](https://docs.briefcaseai.org/features/routing/): Route an AI decision between automatic handling and human review with a narrow, in-process gate. - [Control & Route / Versioned Routing Policy](https://docs.briefcaseai.org/features/versioned-routing-policy/): Reconstruct exactly which rule fired and which policy version was active on any past date. - [Control & Route / Validation Engine](https://docs.briefcaseai.org/features/validation-engine/): Block prompts whose references no longer resolve, before they reach a model. - [Store & Query / Storage Adapters](https://docs.briefcaseai.org/features/storage-adapters/): Choose where your decision records live — the backend that holds your audit trail. - [Store & Query / Bitemporal Storage](https://docs.briefcaseai.org/features/bitemporal-storage/): Answer "what did we know at decision time?" even after a value is later corrected. - [Store & Query / External Data](https://docs.briefcaseai.org/features/external-data/): Pin the exact upstream value a decision used, so it stays reproducible when the source changes. - [Store & Query / RAG Versioning](https://docs.briefcaseai.org/features/rag-versioning/): Tie each retrieval to the exact corpus state, and catch a stale index before it informs a decision. - [Replay & Verify / Deterministic Replay](https://docs.briefcaseai.org/features/replay/): Load a saved decision back through a replay pipeline and compare it against a policy. Re-execution is your executor to supply. - [Replay & Verify / Integrity & Signing](https://docs.briefcaseai.org/features/integrity/): Build verifiable hash chains, canonicalize JSON, and sign records with Ed25519 across Python and TypeScript. - [Replay & Verify / Drift Detection](https://docs.briefcaseai.org/features/drift-detection/): Catch a model whose answers are quietly getting less consistent over time — before it changes what your system decides. - [Replay & Verify / Audit Bundles](https://docs.briefcaseai.org/features/compliance-bundles/): Seal a decision, its evidence, and the policy behind it into one portable, tamper-evident artifact. - [Operate & Evaluate / Cost Tracking](https://docs.briefcaseai.org/features/cost-tracking/): Estimate token costs across platforms and tiers, bill prompt caching, compare models, project monthly spend, and watch budgets — from the data your decisions already carry. - [Operate & Evaluate / OpenTelemetry](https://docs.briefcaseai.org/features/opentelemetry/): Put governed decisions on the same trace timeline as the rest of your stack. - [Operate & Evaluate / Multi-Agent & Events](https://docs.briefcaseai.org/features/multi-agent/): Tie every decision in a multi-step agent pipeline to one workflow and react to them as they happen. - [Operate & Evaluate / Evaluation Runs](https://docs.briefcaseai.org/features/evaluations/): Turn evaluation results into decision records, and replay inspect-ai or lm-eval-harness logs you already have on disk. ## Integrations - [Framework Auto-Instrumentation](https://docs.briefcaseai.org/integrations/auto-instrumentation/): Capture framework decisions globally with reversible instrumentation for seven Python agent frameworks. - [Data & VCS Adapters](https://docs.briefcaseai.org/integrations/data-adapters/): Bind vector retrieval and versioned data reads to exact source commits across supported providers. - [lakeFS](https://docs.briefcaseai.org/integrations/lakefs/): Capture lakeFS commit SHAs for the versioned data your agents read. - [MCP Server](https://docs.briefcaseai.org/integrations/mcp/): Expose Briefcase AI's safe, read-only SDK operations to MCP-capable tools. - [Gymnasium](https://docs.briefcaseai.org/integrations/gymnasium/): Drive a guardrail from a gymnasium.Env, and record any RL rollout as decision records. ## Reference - [Python SDK Guide](https://docs.briefcaseai.org/sdk/python/): Complete guide to the Briefcase AI Python package. - [Python API / Overview](https://docs.briefcaseai.org/api/python/): Full API reference for the Briefcase AI Python SDK, grouped by module. - [Python API / briefcase](https://docs.briefcaseai.org/api/python/briefcase/): Python API reference for briefcase. - [Python API / controls](https://docs.briefcaseai.org/api/python/controls/): Python API reference for gateways, quota ports, provider fallback, throttle classification, and retry helpers. - [Python API / integrity](https://docs.briefcaseai.org/api/python/integrity/): Python API reference for canonical JSON, hash chains, stores, conflict handling, and Ed25519 signatures. - [Python API / auto](https://docs.briefcaseai.org/api/python/auto/): Python API reference for reversible global framework instrumentation. - [Python API / storage](https://docs.briefcaseai.org/api/python/storage/): Python API reference for briefcase.storage. - [Python API / replay](https://docs.briefcaseai.org/api/python/replay/): Python API reference for briefcase.replay. - [Python API / drift](https://docs.briefcaseai.org/api/python/drift/): Python API reference for briefcase.drift. - [Python API / cost](https://docs.briefcaseai.org/api/python/cost/): Python API reference for briefcase.cost. - [Python API / sanitize](https://docs.briefcaseai.org/api/python/sanitize/): Python API reference for briefcase.sanitize. - [Python API / validation](https://docs.briefcaseai.org/api/python/validation/): Python API reference for briefcase.validation. - [Python API / guardrails](https://docs.briefcaseai.org/api/python/guardrails/): Python API reference for briefcase.guardrails. - [Python API / rag](https://docs.briefcaseai.org/api/python/rag/): Python API reference for briefcase.rag. - [Python API / correlation](https://docs.briefcaseai.org/api/python/correlation/): Python API reference for briefcase.correlation. - [Python API / events](https://docs.briefcaseai.org/api/python/events/): Python API reference for briefcase.events. - [Python API / external](https://docs.briefcaseai.org/api/python/external/): Python API reference for briefcase.external. - [Python API / routing](https://docs.briefcaseai.org/api/python/routing/): Python API reference for briefcase.routing. - [Python API / bitemporal](https://docs.briefcaseai.org/api/python/bitemporal/): Python API reference for briefcase.bitemporal. - [Python API / compliance](https://docs.briefcaseai.org/api/python/compliance/): Python API reference for briefcase.compliance. - [Python API / otel](https://docs.briefcaseai.org/api/python/otel/): Python API reference for briefcase.otel. - [Python API / exporters](https://docs.briefcaseai.org/api/python/exporters/): Python API reference for briefcase.exporters. - [Python API / mcp](https://docs.briefcaseai.org/api/python/mcp/): Python API reference for briefcase.mcp. - [Python API / lakefs](https://docs.briefcaseai.org/api/python/lakefs/): Python API reference for briefcase.integrations.lakefs. - [Python API / evals](https://docs.briefcaseai.org/api/python/evals/): Python API reference for briefcase.integrations.evals. - [Python API / gym](https://docs.briefcaseai.org/api/python/gym/): Python API reference for briefcase.integrations.gym. - [TypeScript API / Overview](https://docs.briefcaseai.org/api/typescript/): Package map for the Briefcase AI 4.4 TypeScript controls, native cost engine, and runtime APIs. - [TypeScript API / Controls](https://docs.briefcaseai.org/api/typescript/controls/): TypeScript API reference for invocation controls, ports, retries, pipelines, exporters, and Bedrock helpers. - [TypeScript API / Native Cost Engine](https://docs.briefcaseai.org/api/typescript/native/): TypeScript API reference for the Rust-backed cost calculator and model ID normalization. - [TypeScript API / Runtime](https://docs.briefcaseai.org/api/typescript/runtime/): TypeScript API reference for canonical JSON, hash chains, a typed lakeFS client, connector sync, and agent traces. - [Install & Compatibility](https://docs.briefcaseai.org/reference/install-compat/): Install the oci-bai CLI, bring up the local stack, and check the compatibility matrix. - [Glossary](https://docs.briefcaseai.org/reference/glossary/): Definitions for the core vocabulary used across Briefcase AI — decisions, evidence, policies, bitemporal time, and verifiable bundles. - [Changelog](https://docs.briefcaseai.org/resources/changelog/): Release history for Briefcase AI. - [More / Architecture](https://docs.briefcaseai.org/resources/architecture/): System architecture of Briefcase AI. - [More / Examples](https://docs.briefcaseai.org/resources/examples/): Example code and use cases for Briefcase AI. - [More / FAQ](https://docs.briefcaseai.org/resources/faq/): Frequently asked questions about Briefcase AI. - [More / Artifact Graph](https://docs.briefcaseai.org/evaluate/runs/): How oci-bai tracks every model, fine-tune, and dataset you push — the commit model, provenance, search, and how to compare versions. - [More / Artifact Graph Quick Start](https://docs.briefcaseai.org/evaluate/quickstart/): Push your first tracked image, explore the commit graph, and run your first search in under five minutes. - [More / Artifact Graph CLI](https://docs.briefcaseai.org/evaluate/cli/): Every oci-bai command and its flags. - [More / Artifact Graph Architecture](https://docs.briefcaseai.org/evaluate/architecture/): The gateway, artifact graph, and how pushes become searchable, versioned, deduplicated commits. - [More / Build the Rust Core](https://docs.briefcaseai.org/sdk/rust/): For contributors — build the Briefcase AI Rust core and Python bindings from source. - [More / Development](https://docs.briefcaseai.org/contributing/development/): Set up a development environment for Briefcase AI. - [More / Testing](https://docs.briefcaseai.org/contributing/testing/): Run and write tests for Briefcase AI. - [More / Code Standards](https://docs.briefcaseai.org/contributing/code-standards/): Code style and conventions for Briefcase AI. - [More / Governance](https://docs.briefcaseai.org/contributing/governance/): Project governance and decision-making.