# aCMF

## Docs

- [LLM Roles](https://ecmf.mintlify.app/architecture/llm-roles.md): Adjudicator, Context Enhancer, Deep Memory, and Cortex responsibilities.
- [Memory Lifecycle](https://ecmf.mintlify.app/architecture/memory-lifecycle.md): How memories are created, updated, contradicted, ranked, and snapshot-selected.
- [Architecture Overview](https://ecmf.mintlify.app/architecture/overview.md): Canonical data flow, storage model, and high-level service decomposition.
- [Environment Reference](https://ecmf.mintlify.app/guides/environment-reference.md): Every `ACMF_` environment variable, what it controls, and whether it is required.
- [Operations](https://ecmf.mintlify.app/guides/operations.md): Running aCMF in development, worker schedules, and practical maintenance notes.
- [Provider Configuration](https://ecmf.mintlify.app/guides/providers.md): Configure OpenAI-compatible providers independently for each aCMF role.
- [Retrieval and Scopes](https://ecmf.mintlify.app/guides/retrieval-and-scopes.md): Scope resolution, read modes, and how candidate retrieval works.
- [aCMF](https://ecmf.mintlify.app/index.md): Augmented Cognitive Memory Framework for long-term LLM memory.
- [Quickstart](https://ecmf.mintlify.app/quickstart.md): Run aCMF locally with Docker, migrations, and stub or OpenAI-compatible providers.
- [POST /v1/context](https://ecmf.mintlify.app/reference/context.md): Build prompt-safe grounded context for the next turn.
- [POST /v1/deep-memory](https://ecmf.mintlify.app/reference/deep-memory.md): Answer a grounded memory query with explicit evidence and abstention.
- [POST /v1/process](https://ecmf.mintlify.app/reference/process.md): Queue async post-turn memory adjudication.
- [GET /v1/snapshot/{user_id}](https://ecmf.mintlify.app/reference/snapshot.md): Retrieve the latest hourly user/global snapshot.

## OpenAPI Specs

- [openapi](https://ecmf.mintlify.app/api-reference/openapi.json)