> ## Documentation Index
> Fetch the complete documentation index at: https://docs.ruminate.so/llms.txt
> Use this file to discover all available pages before exploring further.

# A shared project intelligence layer for humans and AI agents.

> Ruminate builds a persistent project brain from your repository. Agents use MCP; humans curate in the web app.

Ruminate builds a persistent **project brain** from your repository — systems, architecture, concepts, and the evidence behind them. Agents read and write that understanding through MCP. Humans inspect, curate, and steer it in the web app.

<Columns cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Sign in, land on your project, connect a repository.
  </Card>

  <Card title="Connect a repository" icon="code-branch" href="/connect">
    Point Ruminate at your own GitHub code.
  </Card>

  <Card title="How it works" icon="layer-group" href="/how-it-works">
    Facts from static analysis, meaning from interpretation.
  </Card>

  <Card title="MCP" icon="plug" href="/mcp">
    Where agents use the project brain.
  </Card>
</Columns>

## What you get

After a repository is analysed you can:

* See the project's current understanding on **Overview**
* Browse **Knowledge** — concepts grounded in real files and line ranges
* Explore **Architecture** — a living map of systems and how they connect
* Track **Decisions**, **Questions**, and **Activity** as the brain grows

Signed-in users land in a default project (named Untitled until you rename it in Settings).

## Two surfaces, one engine

* **MCP** — agents query and write project understanding from the editor
* **Web app** — humans understand, curate, inspect, and steer the brain