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

# Moxn

> Knowledge base for human & agent collaboration — connect via MCP or CLI

Moxn is a knowledge base for human & agent collaboration. Agents connect via MCP or CLI to search, read, and write your docs. Your team edits in a modern web editor. Branching and version control keep everyone in sync — so parallel agents don't clobber each other, and every change is reviewable.

<Frame>
  <img src="https://mintcdn.com/moxn/i8Xr-HNzNiMuDegg/images/kb-landing-page.png?fit=max&auto=format&n=i8Xr-HNzNiMuDegg&q=85&s=27f9e5eecbdf7b9f4ffff85a54c230e8" alt="Moxn knowledge base" width="1512" height="857" data-path="images/kb-landing-page.png" />
</Frame>

***

## The Problem

Four pain points keep showing up when working with agents on documents:

* **Parallel agents, no coordination.** Two agents edit the same doc; the second clobbers the first. No diff, no history, no merge path. Last-write-wins.
* **Markdown in git works for agents, less so for humans.** Agents can navigate a repo, but collaborators lose real-time cursors, rich editing, and comments.
* **MCP bolted onto Notion or Google Docs.** Agents call APIs that weren't designed for them. Writes burn tokens on format translation; search is flat.
* **Permissions.** You want to guard specific content from agent updates or reads without locking it away from humans.

Moxn's approach: a **filesystem** primitive that both humans and agents share, **Git-like branching** so parallel work doesn't deadlock, and **section-grain permissions** that apply independently to humans and agents.

***

## Choose Your Setup

Pick the connection method that fits your workflow:

|                  | CLI                                                 | MCP                                      |
| ---------------- | --------------------------------------------------- | ---------------------------------------- |
| **Auth**         | API key (`MOXN_API_KEY`)                            | Browser-based OAuth                      |
| **Setup**        | `npm install` + `install-skill`                     | Config snippet in MCP settings           |
| **Best for**     | Claude Code, Codex, Cursor, and other coding agents | Claude.ai, ChatGPT, Gemini, desktop apps |
| **How it works** | Agent calls `context` commands directly             | Agent uses MCP tools                     |

<CardGroup cols={2}>
  <Card title="CLI Quick Start" icon="terminal" href="/guides/context-cli">
    API key auth, no MCP needed — ideal for coding agents
  </Card>

  <Card title="MCP Quick Start" icon="plug" href="/quickstart-documents">
    Connect any MCP host — Claude, ChatGPT, Cursor, Gemini, and more
  </Card>
</CardGroup>

***

## Built for Agents

Agents connect to Moxn and get structured access to your content through 13 MCP tools. See the [full MCP reference](/reference/mcp/index).

| Category           | Tools                              |
| ------------------ | ---------------------------------- |
| Discovery          | `find`, `search`, `read`           |
| Content management | `documents`, `sections`            |
| Organization       | `filesystems`, `tags`, `databases` |
| Collaboration      | `merge_requests`, `changesets`     |
| Members & files    | `members`, `upload_url`            |

<CardGroup cols={2}>
  <Card title="Search and read" icon="magnifying-glass">
    Agents query your docs for context, grounding responses in your team's actual knowledge instead of stale training data.
  </Card>

  <Card title="Write on branches" icon="code-branch">
    When you ask an agent to create or update content, it works on a branch. You review the diff and merge when ready.
  </Card>

  <Card title="Control visibility" icon="lock">
    Filesystem roles and section-level AI access controls let you hide sensitive content from agents while keeping it visible to your team.
  </Card>

  <Card title="Any MCP host" icon="plug">
    Claude.ai, ChatGPT, Gemini, Claude Code, Cursor, Claude Desktop, and any other MCP-compatible client.
  </Card>
</CardGroup>

***

## A Real Editor for Your Team

Moxn isn't just an API for agents — it's a full-featured editor for your team. Rich text, code blocks with syntax highlighting, mermaid diagrams, inline comments, and drag-and-drop section reordering.

<Frame>
  <img src="https://mintcdn.com/moxn/i8Xr-HNzNiMuDegg/images/kb-editor.png?fit=max&auto=format&n=i8Xr-HNzNiMuDegg&q=85&s=c39d923d2c7718e8618e142434331533" alt="Moxn document editor with sections, code blocks, and mermaid diagrams" width="1512" height="900" data-path="images/kb-editor.png" />
</Frame>

Your team writes and edits documentation in the web app. Agents read and write the same content over MCP. Everyone works from a single source of truth.

***

## Import & Export

Import from Notion or local files with full fidelity. Work in Moxn with your team and AI agents. Export your changes back when you need to — no lock-in.

For Notion, pick the **in-app integration** (Settings > Integrations — Moxn runs the job for you) or the **CLI** (run yourself, scriptable, CI-friendly). Local files use the CLI. See [Which Path Should I Use?](/migration/index#which-path-should-i-use) for the full breakdown.

<CardGroup cols={2}>
  <Card title="Notion (in-app)" icon="bolt" href="/migration/notion-integration">
    Connect once in Settings > Integrations — no command line needed
  </Card>

  <Card title="Notion (CLI)" icon="terminal" href="/migration/notion">
    Same import, run from your terminal or CI
  </Card>

  <Card title="Import Local Files" icon="file-import" href="/migration/local">
    Markdown, text, and CSV from your filesystem
  </Card>

  <Card title="Export to Local Files" icon="file-export" href="/migration/export-local">
    Download as markdown with images, PDFs, and CSVs
  </Card>
</CardGroup>