# Docs

> Documentation for the Engine.

## Docs

- [Authentication](https://internal.september.wtf/bap-engine/api-reference/auth.md): How bap-engine authenticates products and admins.
- [Admin endpoints](https://internal.september.wtf/bap-engine/api-reference/endpoints/admin.md): Register products and update policies. Requires X-Admin-Key.
- [Discovery and admission endpoints](https://internal.september.wtf/bap-engine/api-reference/endpoints/discovery.md): Look up an engine, or admit a request with auto-provision and auto-wake.
- [Fleet endpoints](https://internal.september.wtf/bap-engine/api-reference/endpoints/fleet.md): List engines, fleet status, and rolling metrics.
- [Health endpoint](https://internal.september.wtf/bap-engine/api-reference/endpoints/health.md): The orchestrator's own liveness check.
- [Lifecycle endpoints](https://internal.september.wtf/bap-engine/api-reference/endpoints/lifecycle.md): Provision, start, stop, destroy, and rotate-key for one user's engine.
- [Errors](https://internal.september.wtf/bap-engine/api-reference/errors.md): Every error code bap-engine returns and how to handle each.
- [API reference](https://internal.september.wtf/bap-engine/api-reference/overview.md): Every endpoint bap-engine exposes.
- [Architecture](https://internal.september.wtf/bap-engine/architecture.md): What's inside bap-engine — the orchestrator, registry, lifecycle, health loop, and backends.
- [Audit log](https://internal.september.wtf/bap-engine/audit.md): Every lifecycle action the orchestrator takes is recorded. How to read the table and what it answers.
- [Backends](https://internal.september.wtf/bap-engine/backends.md): How the orchestrator actually launches engine processes — Docker (production) and subprocess (dev).
- [Environment variables](https://internal.september.wtf/bap-engine/config/env-vars.md): Every environment variable bap-engine reads.
- [Production deploy](https://internal.september.wtf/bap-engine/deploy.md): Bring bap-engine up in production — secrets, TLS, backups, observability.
- [Engine contract](https://internal.september.wtf/bap-engine/engine-contract.md): What the orchestrator promises engines and what engines promise back.
- [Health and auto-restart](https://internal.september.wtf/bap-engine/health.md): How the orchestrator probes engines, detects failures, and restarts with backoff.
- [Engine lifecycle](https://internal.september.wtf/bap-engine/lifecycle.md): The state machine the orchestrator drives for every engine — provision, run, sleep, wake, stop, fail, destroy.
- [BAP Engine](https://internal.september.wtf/bap-engine/overview.md): The orchestrator that runs and manages a fleet of Engine instances per user.
- [Policy](https://internal.september.wtf/bap-engine/policy.md): Quota, rate-limit, and admission control inside the orchestrator.
- [Quickstart](https://internal.september.wtf/bap-engine/quickstart.md): Run bap-engine locally with Postgres + the orchestrator + a test Engine in under twenty minutes.
- [Security](https://internal.september.wtf/bap-engine/security.md): Keys, encryption, rotation, and the trust boundaries inside bap-engine.
- [Components](https://internal.september.wtf/internal/architecture/components.md): C4 Level 3: the major components inside the Engine and what each one is responsible for.
- [Containers](https://internal.september.wtf/internal/architecture/containers.md): C4 Level 2: the deployable units of the Engine and how they talk to each other.
- [Data flow](https://internal.september.wtf/internal/architecture/data-flow.md): What calls what during a single /execute request, end to end.
- [Architecture overview](https://internal.september.wtf/internal/architecture/overview.md): How the Engine is built — the one-page mental model every engineer should hold in their head.
- [Shared types](https://internal.september.wtf/internal/architecture/types.md): The contract types that cross module boundaries inside the Engine.
- [Contributing](https://internal.september.wtf/internal/contributing.md): How to add or change docs in this repo.
- [Glossary](https://internal.september.wtf/internal/glossary.md): One canonical name for each concept in the Engine.
- [Style guide](https://internal.september.wtf/internal/style-guide.md): How we write docs.
- [Testing strategy](https://internal.september.wtf/internal/testing/strategy.md): How we test the Engine — unit, integration, end-to-end, and the rules that keep tests honest.
- [Environment variables](https://internal.september.wtf/operations/config/env-vars.md): Every environment variable the Engine reads, what it does, and what it defaults to.
- [Feature flags](https://internal.september.wtf/operations/config/feature-flags.md): All Engine feature flags grouped by area, with rollout history and recommended defaults.
- [Secrets](https://internal.september.wtf/operations/config/secrets.md): Where secrets live, how they get to the Engine, and how to rotate them.
- [Deployment overview](https://internal.september.wtf/operations/deploy/overview.md): How the Engine deploys, and which topology matches your needs.
- [Production deploy](https://internal.september.wtf/operations/deploy/prod-deploy.md): Step-by-step bring-up of the Engine in production.
- [Rollback](https://internal.september.wtf/operations/deploy/rollback.md): Undoing a bad deploy quickly and cleanly.
- [Upgrade](https://internal.september.wtf/operations/deploy/upgrade.md): Moving an existing Engine deployment to a new version.
- [Database](https://internal.september.wtf/operations/infra/database.md): Operating the brain SQLite database.
- [Observability](https://internal.september.wtf/operations/infra/observability.md): Logs, metrics, traces, and dashboards for a production Engine.
- [Common tasks](https://internal.september.wtf/operations/local-dev/common-tasks.md): Useful one-liners for local Engine development.
- [Prerequisites](https://internal.september.wtf/operations/local-dev/prerequisites.md): What you need installed before running the Engine locally.
- [Local development quickstart](https://internal.september.wtf/operations/local-dev/quickstart.md): Run the Engine on your laptop in under fifteen minutes.
- [Troubleshooting local dev](https://internal.september.wtf/operations/local-dev/troubleshooting.md): Symptoms, causes, and fixes for common problems running the Engine locally.
- [On-call handoff](https://internal.september.wtf/operations/on-call/handoff.md): What gets passed from the off-going to the in-coming on-call.
- [On-call rotation](https://internal.september.wtf/operations/on-call/rotation.md): How the on-call rotation runs.
- [Severity definitions](https://internal.september.wtf/operations/on-call/severity.md): What Sev1, Sev2, and Sev3 mean for the Engine.
- [Postmortems](https://internal.september.wtf/operations/postmortems/overview.md): Blameless incident reviews. How we run them and how we file them.
- [Access control](https://internal.september.wtf/operations/security/access-control.md): Who can do what against an Engine.
- [Data handling](https://internal.september.wtf/operations/security/data-handling.md): What data the Engine stores, where, and for how long.
- [Threat model](https://internal.september.wtf/operations/security/threat-model.md): What we defend against and how.
- [SLO definitions](https://internal.september.wtf/operations/slo/definitions.md): Service level objectives for the Engine.
- [The agent loop](https://internal.september.wtf/platform/agents-and-tools/agent-loop.md): How the Engine drives an agent from message to answer.
- [File search](https://internal.september.wtf/platform/agents-and-tools/built-in-tools/file-search.md): Searching across files inside the Engine's sandbox.
- [MCP connectors](https://internal.september.wtf/platform/agents-and-tools/built-in-tools/mcp.md): Connecting external services via the Model Context Protocol so the agent can act outside the sandbox.
- [Web search](https://internal.september.wtf/platform/agents-and-tools/built-in-tools/web-search.md): Built-in web search via Brave Search and Tavily fallback.
- [Defining tools](https://internal.september.wtf/platform/agents-and-tools/defining-tools.md): How to add a tool the agent can call — schemas, registration, lifecycle.
- [Durability and resume](https://internal.september.wtf/platform/agents-and-tools/durability.md): How tasks survive disconnects, restarts, and HITL pauses.
- [Multi-agent patterns](https://internal.september.wtf/platform/agents-and-tools/multi-agent.md): Coordinating multiple agents within one task — when to compose, when to keep separate.
- [Permissions](https://internal.september.wtf/platform/agents-and-tools/permissions.md): How the Engine decides which tool calls require explicit user approval, and how to configure it.
- [Authentication](https://internal.september.wtf/platform/api-reference/auth.md): How the Engine authenticates HTTP requests.
- [Admin endpoints](https://internal.september.wtf/platform/api-reference/endpoints/admin.md): Operational endpoints for reloading catalogs and triggering background work.
- [Asset Directory endpoints](https://internal.september.wtf/platform/api-reference/endpoints/assets.md): Manage MCP server connections — list, connect, disconnect, OAuth callbacks.
- [POST /execute](https://internal.september.wtf/platform/api-reference/endpoints/execute.md): Run an agent turn against a task. Streams Server-Sent Events.
- [HITL endpoints](https://internal.september.wtf/platform/api-reference/endpoints/hitl.md): Endpoints for answering questions the agent asks during execution.
- [Memory endpoints](https://internal.september.wtf/platform/api-reference/endpoints/memory.md): Search, export, and import the Engine's long-term memory.
- [Observability endpoints](https://internal.september.wtf/platform/api-reference/endpoints/observability.md): Health and feedback endpoints.
- [POST /rpc](https://internal.september.wtf/platform/api-reference/endpoints/rpc.md): Generic dispatcher for skill and tool invocation outside the agent loop.
- [Skills endpoints](https://internal.september.wtf/platform/api-reference/endpoints/skills.md): List, create, update, and delete custom skills.
- [Errors](https://internal.september.wtf/platform/api-reference/errors.md): Engine error codes, when you'll see them, and how to handle them.
- [API reference](https://internal.september.wtf/platform/api-reference/overview.md): The complete HTTP API surface of the Engine, by category.
- [Streaming events](https://internal.september.wtf/platform/api-reference/streaming-events.md): Every Server-Sent Event the Engine can emit during /execute.
- [Context strategy](https://internal.september.wtf/platform/build-with-engine/context-strategy.md): Designing what the model sees on each turn so the agent stays sharp without burning tokens.
- [Error handling](https://internal.september.wtf/platform/build-with-engine/error-handling.md): Patterns for retries, fallbacks, and graceful degradation when things go wrong.
- [Evaluation](https://internal.september.wtf/platform/build-with-engine/evaluation.md): How to know whether the agent you built actually works, and stays working.
- [Prompt engineering](https://internal.september.wtf/platform/build-with-engine/prompt-engineering.md): Patterns that work for the system prompt and tool descriptions in an agent built on the Engine.
- [Rate limits](https://internal.september.wtf/platform/build-with-engine/rate-limits.md): Where rate limits come from and how to live within them.
- [Safety](https://internal.september.wtf/platform/build-with-engine/safety.md): Building agents that don't cause harm — to your users, to your systems, or to anyone downstream.
- [Code execution](https://internal.september.wtf/platform/capabilities/code-execution.md): Running code inside the Engine's sandbox during a turn.
- [Multimodal](https://internal.september.wtf/platform/capabilities/multimodal.md): Sending images, files, and other media through the Engine.
- [Streaming](https://internal.september.wtf/platform/capabilities/streaming.md): Reading the SSE stream from /execute, with patterns for resilient client-side handling.
- [Structured outputs](https://internal.september.wtf/platform/capabilities/structured-outputs.md): How to get JSON back from the agent that matches a schema.
- [Text generation](https://internal.september.wtf/platform/capabilities/text-generation.md): How the Engine generates text and what knobs you have to control it.
- [Tool use](https://internal.september.wtf/platform/capabilities/tool-use.md): How the agent calls tools and how to define them.
- [Libraries](https://internal.september.wtf/platform/get-started/libraries.md): How to talk to the Engine from your application.
- [Models and pricing](https://internal.september.wtf/platform/get-started/models-and-pricing.md): Models the Engine supports and what they cost.
- [Welcome to the Engine](https://internal.september.wtf/platform/get-started/overview.md): An introduction to the Engine — what it is, what it's for, and how to think about building with it.
- [Quickstart](https://internal.september.wtf/platform/get-started/quickstart.md): Make your first request to the Engine — in curl, Python, or Node.
- [Quickstart with curl](https://internal.september.wtf/platform/get-started/quickstart-curl.md): Make your first Engine call from the command line, no dependencies needed.
- [Quickstart with Node](https://internal.september.wtf/platform/get-started/quickstart-node.md): Make your first Engine call from Node.js or TypeScript.
- [Quickstart with Python](https://internal.september.wtf/platform/get-started/quickstart-python.md): Make your first Engine call from Python.
- [Build a coding agent](https://internal.september.wtf/platform/guides/build-a-coding-agent.md): Walk through building an agent that reads a codebase, makes a change, runs tests, and reports back.
- [Build a research agent](https://internal.september.wtf/platform/guides/build-a-research-agent.md): Walk through building an agent that researches a topic, synthesizes findings, and produces a cited summary.
- [Build a customer support agent](https://internal.september.wtf/platform/guides/build-a-support-agent.md): Walk through building an agent that handles support tickets, escalates when needed, and learns from feedback.
- [Changelog](https://internal.september.wtf/platform/resources/changelog.md): Engine release history.
- [FAQ](https://internal.september.wtf/platform/resources/faq.md): Common questions about building on the Engine.
- [Glossary](https://internal.september.wtf/platform/resources/glossary.md): Plain-English definitions of every Engine concept.
- [Migration guides](https://internal.september.wtf/platform/resources/migration-guides.md): Step-by-step upgrades across Engine versions.
- [Eval harness](https://internal.september.wtf/platform/test-and-evaluate/eval-harness.md): Setting up an automated evaluation runner against an Engine deployment.
- [Golden datasets](https://internal.september.wtf/platform/test-and-evaluate/golden-datasets.md): Building and curating the case set your evals run against.
- [Observability](https://internal.september.wtf/platform/test-and-evaluate/observability.md): What the Engine emits and how to consume it from your application.
- [Regression testing](https://internal.september.wtf/platform/test-and-evaluate/regression.md): Catching agent regressions when you change the model, the prompt, the catalog, or the Engine.