> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.livingbrain.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.livingbrain.com/_mcp/server.

# Living Brain API

> Build with Living Brain: an API, SDKs, CLI, and MCP server that turn what your users read, write, and say into a knowledge vault that compounds.

Living Brain API

Now in early access

# Never start from zero again.

Living Brain turns everything your users read, write, and say into a knowledge vault
that compounds — surfacing non-obvious connections across conversations, documents,
meetings, and time.

[Quickstart](/docs/quickstart)

[API Reference](/api-reference)

```bash
# Authenticate and create a brain
$ curl https://api.livingbrain.com/v1/brains \
  -H "x-api-key: $LIVING_BRAIN_API_KEY" \
  -H "x-subject-id: user_123" \
  --json '{"name": "Research"}'
```

```json
# 201 Created
{ "id": "brn_01", "name": "Research",
  "slug": "research", "subjectId": "user_123" }
```

MCP server

## Give your agent a brain.

Every Living Brain operation is a tool on the hosted MCP server. Paste one snippet into
Claude Code, Codex, Cursor, or VS Code and your agent can capture, search, and recall
knowledge across sessions.

[Read the MCP guide →](/mcp)

[**Using ChatGPT or Claude?**Connect Living Brain with your account sign-in. No API key or code. →](/mcp/sign-in)

**Connect your MCP client** · [Browse all 60 tools ↗](https://livingbrain.fernmcp.dev/livingbrain)

**Claude Code** (Terminal): Run it in a terminal. Claude Code picks the server up on its next start.

```bash
claude mcp add --transport http livingbrain \
  https://livingbrain.fernmcp.dev/livingbrain \
  --header 'x-api-key: YOUR_API_KEY' \
  --header 'x-subject-id: YOUR_SUBJECT_ID'
```

**Codex** (\~/.codex/config.toml): A url key rather than a command is what makes it remote. codex mcp add --url writes the entry, but not the headers.

```toml
[mcp_servers."livingbrain"]
url = "https://livingbrain.fernmcp.dev/livingbrain"
http_headers = { "x-api-key" = "YOUR_API_KEY", "x-subject-id" = "YOUR_SUBJECT_ID" }
```

**Cursor** (\~/.cursor/mcp.json): Global to every project, or .cursor/mcp.json for one. Cursor lists them under Customize in the sidebar.

```json
{
  "mcpServers": {
    "livingbrain": {
      "type": "http",
      "url": "https://livingbrain.fernmcp.dev/livingbrain",
      "headers": {
        "x-api-key": "YOUR_API_KEY",
        "x-subject-id": "YOUR_SUBJECT_ID"
      }
    }
  }
}
```

**VS Code** (.vscode/mcp.json): This file is committed with the project, so put the credential in an inputs prompt rather than in the file.

```json
{
  "servers": {
    "livingbrain": {
      "type": "http",
      "url": "https://livingbrain.fernmcp.dev/livingbrain",
      "headers": {
        "x-api-key": "YOUR_API_KEY",
        "x-subject-id": "YOUR_SUBJECT_ID"
      }
    }
  }
}
```

**Other**: The endpoint and the headers to send with it, for any other client that supports remote HTTP servers.

```text
https://livingbrain.fernmcp.dev/livingbrain
x-api-key: YOUR_API_KEY
x-subject-id: YOUR_SUBJECT_ID
```

For Claude Code, add `--scope user` to install the server for every project instead of only the current one.

CLI

## Or let it use the terminal.

Any coding agent that can run shell commands can use the `livingbrain` CLI directly.
Install it, set your key, and ask the agent to save or find something.

[Read the CLI guide →](/cli)

Example: a coding agent saving a note with the CLI.

> Save the notes from today's onboarding call to my Research brain

The agent runs:

```bash
livingbrain brains list --query "[?name=='Research'].id"
```

```json
["brn_01"]
```

```bash
livingbrain captures capture --brain-id brn_01 --json '{"kind": "note", "content": "The customer prefers weekly check-ins and dislikes long emails."}'
```

```json
{ "id": "src_01", "kind": "note", "status": "pending" }
```

Saved to Research. Living Brain links it to related pages once it finishes compiling.

## Start here

#### [Guides](/docs/quickstart)

Quickstart, authentication, core concepts, and how to capture, search, and sync a brain.

#### [SDKs](/sdk/typescript)

Typed clients for TypeScript and Python, so you can call the API without hand-rolling requests.

#### [CLI](/cli)

Call every endpoint from your terminal or CI with the `livingbrain` command.

#### [MCP](/mcp)

Connect Claude Code, Cursor, or VS Code to the hosted MCP server and use a brain as agent memory.