# Brain Framework > 🧠 Brain Framework: from information to informed actions. Brain Framework (`bf`) keeps a brain in plain files: collection, validation, a knowledge graph and offline retrieval for people and agents. Release notes and upgrade steps are in the [changelog](https://raw.githubusercontent.com/fmind/brain-framework/main/CHANGELOG.md). ## Start here - [Overview](): Keep evidence, decisions and next steps in plain files that you and your agents search, read and connect. - [Getting started](): Create a brain, save your first project decision, search its reason and verify the result. - [Four-tool walkthrough](): Run a fictional four-tool example that collects evidence, records a conclusion and flags it when a source changes. - [Core concepts](): Understand projects, records, concepts and actions, the flow from evidence to decisions, and the terms BF uses. - [Install and update](): Install or update Brain Framework, match documentation to your version and verify the result. ## Guides - [Write notes and actions](): Write project notes, reusable concepts and resumable actions with complete Markdown examples. - [Search and read](): Find saved evidence, read exact sources and check whether results are complete. - [Linking knowledge](): Connect notes and records with citations, typed fields, explicit identities and declared relations. - [Check your brain](): Validate saved files and test that recurring questions still retrieve the expected evidence. - [Add a sensor](): Collect a selected local document or highlight, configure sensors and understand update and replacement rules. - [Run routines](): Run deterministic brain programs now, on a schedule or from Git hooks, and turn their output into logs or review actions. - [Run prompts](): Run your agent harness over the brain on a schedule or on demand, and keep every run as a record it reads back. - [Watch and schedule updates](): Run or observe due sensors, routines and prompts, set watch preferences and generate native schedules. - [Agent workflows](): Use the CLI, the packaged skills and optional hooks to find evidence and keep agent work resumable. - [Connect an agent with MCP](): Connect an agent through local MCP search and read tools, then verify access to a saved decision. - [Team setup](): Set up a private team brain, share reviewed evidence, coordinate collection and evaluate a pilot. ## Reference - [Commands](): Look up Brain Framework commands, previews, graph exports, replies, errors and exit codes. - [Configuration](): Configure a brain, choose which brains a command uses, declare related brains and locate local state. - [Notes, records and fields](): Exact formats of bf.yaml, notes, records, shared fields and sensor mappings, with editor validation. - [Links and tags](): Exact rules for BF addresses, section anchors, identities, relation links and tags. - [Retrieval](): Reply fields and rules for search, read, pages, relation pages, continuations and MCP clients. - [Limits and safeguards](): Look up size, page and process limits, and the safeguards around running programs. - [Troubleshooting](): Diagnose installation, retrieval, collection and file problems without losing evidence. ## Privacy and legal - [Privacy and security](): Understand local storage, cloud-agent boundaries, program execution and audience separation.