# Rabbit Brain docs

> Rabbit Brain (rb) keeps the record of the ML experiments you run with coding agents, as plain YAML in .rb/ next to your code. If a project has a .rb/ directory, run `rb context --model <your model id>` before anything else: it is the state of the work. The full manual is built into rb: `rb docs`.

Every page below but the Changelog as one file: https://rabbitbrain.ai/docs/llms-full.txt

## Pages

- [Rabbit Brain docs](https://rabbitbrain.ai/docs/): Rabbit Brain (rb) keeps the record of the ML experiments you run with coding agents, as plain YAML in .rb/ next to your code.
- [Install](https://rabbitbrain.ai/docs/install/): Install rb with one command, check it, and take the one-minute tour.
- [Your first record](https://rabbitbrain.ai/docs/first-record/): Start the record in your repository, paste one line into your coding agent, and read where the work stands.
- [Connect your agent](https://rabbitbrain.ai/docs/connect/): What rb installs for Claude Code, Codex and Cursor, its session hook, how to add it later, and how to use rb from a plain shell.
- [The record](https://rabbitbrain.ai/docs/concepts/record/): Plain YAML files in .rb/, committed with your code. What the work asks, how each question is tested, what ran, and the calls made on it.
- [Criteria](https://rabbitbrain.ai/docs/concepts/criteria/): A criterion is a rule over one comparison, declared by someone at a time. rb reads every finding against one.
- [Findings](https://rabbitbrain.ai/docs/concepts/findings/): rb computes each criterion's finding from the runs every time the record is read, and says it on separate axes, never in one word.
- [Calls](https://rabbitbrain.ai/docs/concepts/calls/): A call accepts or rejects a result, or moves a line of work on. Calls are yours; one an agent records waits until you confirm it.
- [Sharing](https://rabbitbrain.ai/docs/concepts/sharing/): Send one page, serve the record on your machine, publish a bundle anyone can clone and rerun, and what rabbitbrain.ai serves.
- [Cookbooks](https://rabbitbrain.ai/docs/cookbooks/): Eleven tasks, start to finish. Each has the prompt to paste into your coding agent, and rb's real output from running it.
- [Start tracking a project already under way](https://rabbitbrain.ai/docs/cookbooks/project-under-way/): Record where the work stands from the repository's own files, in one turn, with no questions asked first.
- [Record one table of many runs](https://rabbitbrain.ai/docs/cookbooks/one-table/): A results.tsv in the top folder, one row a run. rb reconcile names it, and one rb run add records every row.
- [Record a folder per run](https://rabbitbrain.ai/docs/cookbooks/run-folders/): Each run keeps its own folder with its config and metrics. rb reconcile says what each config records, and prints the rb run add line for each folder.
- [Check a summary's claims against the runs](https://rabbitbrain.ai/docs/cookbooks/check-a-summary/): Turn each claim in an agent's summary into the criterion it states, cited to its line, and let rb read it from the runs.
- [Declare the criterion before the runs](https://rabbitbrain.ai/docs/cookbooks/criterion-before-runs/): Declare what counts and freeze it before the first run, so every finding says the criterion came first.
- [Compare against a published baseline](https://rabbitbrain.ai/docs/cookbooks/published-baseline/): Cite the paper's number and the settings it used. rb says reproduced or not reproduced, and does not compare when a setting differs.
- [Get an interval, not a single number](https://rabbitbrain.ai/docs/cookbooks/interval-from-seeds/): Run seeds until the criterion has the interval it needs, and read what rb says after each one.
- [Read a rule over seeds](https://rabbitbrain.ai/docs/cookbooks/rules-over-seeds/): A count, a best of N and a relative change over two arms and five seeds, each read its own way.
- [Share the record, and rerun it](https://rabbitbrain.ai/docs/cookbooks/share-a-record/): Write a page your advisor reads, publish a bundle with a Rerun this research prompt, and watch a reader's agent clone it and rerun it.
- [Rerun a run from its receipt](https://rabbitbrain.ai/docs/cookbooks/rerun-from-receipt/): The receipt names the command and the commit. Run it again on that code, and record the rerun as a check of the run.
- [rb with Claude Code, Codex and Cursor](https://rabbitbrain.ai/docs/cookbooks/three-agents/): One record, whichever agent writes to it. What rb installs for each, and how a running session differs from your terminal.
- [The CLI](https://rabbitbrain.ai/docs/reference/cli/): Every rb command, with its help exactly as rb prints it.
- [Commands: The record](https://rabbitbrain.ai/docs/reference/cli/the-record/): These commands start the record and write to it: the project's title, questions, experiments, variants, notes, metrics, settings, criteria, runs, analyses and resources.
- [Commands: Calls](https://rabbitbrain.ai/docs/reference/cli/calls/): These commands lock an experiment before its evidence, record a call, confirm a call or a design in your own terminal, and take anything back.
- [Commands: Reading it](https://rabbitbrain.ai/docs/reference/cli/reading/): These commands read the record: in your browser, in your terminal, as one briefing for an agent, or as one page to send.
- [Commands: Carrying it](https://rabbitbrain.ai/docs/reference/cli/carrying/): These commands take the record somewhere else: an export with a manifest, a bundle another repository clones, and a check of this machine against the runs.
- [Commands: For agents](https://rabbitbrain.ai/docs/reference/cli/for-agents/): These commands are what a coding agent uses: the MCP server, the skill, the journal, reconcile and compile.
- [Commands: Upkeep](https://rabbitbrain.ai/docs/reference/cli/upkeep/): These commands keep the record sound: check it, write its files in canonical form, and rb's own settings and optional usage analytics.
- [Commands: Leaving](https://rabbitbrain.ai/docs/reference/cli/leaving/): This command takes rb out of a checkout, and only what checks out as rb's.
- [Commands: Reference](https://rabbitbrain.ai/docs/reference/cli/reference/): These commands are the tour, the manual, the record's format, the version and the feedback link.
- [Commands: rabbitbrain.ai](https://rabbitbrain.ai/docs/reference/cli/rabbitbrain-ai/): These commands sign a machine in to an account at rabbitbrain.ai, keep a private copy of a record's page there, and revoke the machine's credential. Anyone can make an account, at rabbitbrain.ai/signin.
- [The record's format](https://rabbitbrain.ai/docs/reference/schema/): Rabbit Brain formats 1 and 2, one JSON Schema for every file in .rb/, and rb's vocabulary of record kinds.
- [Changelog](https://rabbitbrain.ai/docs/changelog/): What each release of rb changed, newest first, from rb's own CHANGELOG.md.
- [llms.txt](https://rabbitbrain.ai/docs/agents/llms-txt/): Where an agent finds rb's documentation: these docs as plain text, the manual built into rb, and what an agent must not do.
- [Auth](https://rabbitbrain.ai/docs/agents/auth/): How rb signs a machine in to rabbitbrain.ai or another hub, where the credential lives, and what an agent may and may not touch.