@m4ykeldev/notebooklm-mcp

v0.2.6

Published

11 days ago

MCP server for Google NotebookLM — 32 tools for notebooks, sources, research, and studio content generation

0High
0Medium
0Low

m4ykeldev

mcp notebooklm google ai model-context-protocol

🧠 NotebookLM MCP Server

Bridge the Gap Between Google NotebookLM and Your AI Workspace

Stop jumping between browser tabs. NotebookLM MCP brings the full analytical power of Google NotebookLM directly into your local terminal, IDE, and AI assistants like Claude, Cursor, and VS Code.

Manage notebooks, ingest diverse sources, trigger deep research, and generate studio-quality content—all via a single, standardized Model Context Protocol (MCP) interface.

New to MCP? The Model Context Protocol is a standard for connecting LLMs to external data sources and tools. This package speaks MCP over stdio — your AI client (Claude Desktop, Cursor, VS Code, etc.) spawns notebooklm-mcp serve as a subprocess and the two communicate over JSON-RPC. The 32 tools below become callable functions in the model's tool list.

🔥 Key Capabilities

⚡ Seamless Authentication: Log in once with notebooklm-mcp auth. Our automated CDP-based flow handles secure cookie extraction so you can focus on your data.
🔄 Resilient Connectivity: Built-in background session restoration. If your session expires, the server transparently reconnects without breaking your workflow.
📂 Universal Ingestion: Instantly add URLs, YouTube transcripts, Google Drive files, or raw text snippets to any notebook.
🕵️ Autonomous Research: Harness Google's Deep Research engine. Start a task, poll its progress, and import structured insights directly into your project.
🎭 Creative Studio: Programmatically generate Audio Overviews (podcasts), Briefing Docs, Infographics, Slide Decks, and Quizzes from your sources.

🚀 Quick Start

1. Installation

Run it instantly with npx:

npx -y @m4ykeldev/notebooklm-mcp serve

Or install globally for better performance:

npm install -g @m4ykeldev/notebooklm-mcp

Developers who want to hack on the source: this repo uses pnpm (pinned via packageManager in package.json). After cloning, run corepack enable && pnpm install. See CONTRIBUTING.md on GitHub for the full dev / release flow.

2. The "One-Click" Login

Say goodbye to manual cookie hunting. Our smart auth flow does the heavy lifting for you.

notebooklm-mcp auth

A secure Chrome window will open. Simply log into your Google account, and we'll handle the rest. Your session is stored locally and securely.

Auth fallbacks if automated Chrome can't run:

notebooklm-mcp auth --manual              # interactive copy/paste from your browser
notebooklm-mcp auth --file tokens.json    # import a previously exported bundle
notebooklm-mcp auth --show-tokens         # verify the cached session

For headless / CI environments, set NOTEBOOKLM_COOKIES (and optionally NOTEBOOKLM_CSRF_TOKEN, NOTEBOOKLM_SESSION_ID) instead of running the auth flow. Token resolution order: env var → ~/.notebooklm-mcp/auth.json → error.

🤖 AI Assistant Integration

Pick your client below — every section shows the config file path per OS, a minimal copy-paste block, and the advanced variant with --query-timeout and NOTEBOOKLM_COOKIES env override.

| Client | Config file | Format | |---|---|---| | Claude Desktop | claude_desktop_config.json | JSON | | Claude Code (CLI) | ~/.claude.json or claude mcp add | JSON / CLI | | Codex CLI | ~/.codex/config.toml | TOML | | OpenAI Agents SDK (Python) | in-code | Python | | OpenAI Agents SDK (TypeScript) | in-code | TypeScript | | Gemini CLI | ~/.gemini/settings.json | JSON | | Cursor | ~/.cursor/mcp.json | JSON | | VS Code (Copilot Chat agent mode) | .vscode/mcp.json | JSON | | Windsurf | ~/.codeium/windsurf/mcp_config.json | JSON | | JetBrains AI Assistant / Junie | ~/.junie/mcp/mcp.json | JSON | | Zed | ~/.config/zed/settings.json | JSON | | OpenCode (sst) | opencode.jsonc | JSONC | | Cline (VS Code extension) | extension settings UI | JSON | | Goose (Block) | ~/.config/goose/config.yaml | YAML | | 5ire | in-app settings | GUI | | Aider | not yet supported | — | | Generic stdio caller | yours | — |

Claude Desktop

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Linux: ~/.config/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
    }
  }
}

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000", "--debug"],
      "env": {
        "NOTEBOOKLM_COOKIES": "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..."
      }
    }
  }
}

Verify: restart Claude Desktop, click the hammer icon at the bottom-right of the input. Gotchas: absolute paths only; relative paths fail silently on startup.

Claude Code (CLI)

Recommended path is the claude mcp add CLI (writes the JSON for you):

claude mcp add --transport stdio notebooklm -- npx -y @m4ykeldev/notebooklm-mcp serve

{
  "mcpServers": {
    "notebooklm": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
    }
  }
}

claude mcp add --transport stdio \
  --env NOTEBOOKLM_COOKIES="SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..." \
  notebooklm -- npx -y @m4ykeldev/notebooklm-mcp serve --query-timeout 180000

Verify: claude mcp list && claude mcp get notebooklm, or /mcp inside a session. Gotchas: all options before the server name; use -- to separate options from the command. Project-scoped .mcp.json needs interactive approval on first sight.

Codex CLI

macOS: ~/.codex/config.toml
Linux: ~/.config/codex/config.toml
Windows: %APPDATA%\codex\config.toml

[mcp_servers.notebooklm]
command = "npx"
args = ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]

[mcp_servers.notebooklm]
command = "npx"
args = ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"]
env = { NOTEBOOKLM_COOKIES = "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..." }

Verify: codex --list-tools should list the NotebookLM tools. Gotchas: Codex CLI's MCP schema is still being formalized — double-check against the latest openai/codex README.

OpenAI Agents SDK — Python

Wire it in code (no config file):

from agents.mcp import MCPServerStdio

async with MCPServerStdio(
    name="NotebookLM",
    params={
        "command": "npx",
        "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"],
    },
) as server:
    tools = await server.list_tools()
    print([t.name for t in tools])  # 32 tools

async with MCPServerStdio(
    name="NotebookLM",
    params={
        "command": "npx",
        "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"],
        "env": {"NOTEBOOKLM_COOKIES": "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..."},
    },
) as server:
    ...

OpenAI Agents SDK — TypeScript

import { MCPServerStdio } from "@openai/agents";

const server = new MCPServerStdio({
  command: "npx",
  args: ["-y", "@m4ykeldev/notebooklm-mcp", "serve"],
});

await server.connect();
const tools = await server.listTools();
console.log(tools.map((t) => t.name)); // 32 tools

const server = new MCPServerStdio({
  command: "npx",
  args: ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"],
  env: { NOTEBOOKLM_COOKIES: "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..." },
});

Gemini CLI

~/.gemini/settings.json (user-global) or .gemini/settings.json (per-project).

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
    }
  }
}

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"],
      "timeout": 180000,
      "env": {
        "NOTEBOOKLM_COOKIES": "$NOTEBOOKLM_COOKIES"
      }
    }
  }
}

Verify: gemini mcp list, or /mcp in a session. Gotchas: undefined $VAR resolves to empty string — pre-export them in your shell.

Cursor

macOS / Linux: ~/.cursor/mcp.json (global) or .cursor/mcp.json (per-project)
Windows: %APPDATA%\Cursor\mcp.json

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
    }
  }
}

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"],
      "env": {
        "NOTEBOOKLM_COOKIES": "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..."
      }
    }
  }
}

Verify: Cursor Settings → MCP should show notebooklm in green, or MCP: View Server Status in the Command Palette. Gotchas: Cursor only loads MCP servers at startup — fully quit and relaunch after edits. Soft ~40-tool ceiling across all enabled servers combined.

VS Code (Copilot Chat agent mode)

.vscode/mcp.json (per-project). Top-level key is servers (not mcpServers — Microsoft renamed this).

{
  "servers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
    }
  }
}

{
  "servers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"],
      "env": {
        "NOTEBOOKLM_COOKIES": "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..."
      }
    }
  }
}

Verify: Command Palette → MCP: Open User Configuration. Ask Copilot Chat in agent mode to list NotebookLM notebooks.

Windsurf

macOS / Linux: ~/.codeium/windsurf/mcp_config.json
Windows: %APPDATA%\Codeium\Windsurf\mcp_config.json

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
    }
  }
}

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"],
      "env": {
        "NOTEBOOKLM_COOKIES": "${env:NOTEBOOKLM_COOKIES}"
      }
    }
  }
}

Verify: open the Cascade sidebar → MCP settings → confirm notebooklm is active. Gotchas: prefer ${env:VAR} / ${file:/path} interpolation over inlining cookies; Windsurf substitutes at server-launch time so secrets stay out of the config file.

JetBrains AI Assistant / Junie

macOS / Linux: ~/.junie/mcp/mcp.json (user-global) or .junie/mcp/mcp.json (per-project)
Windows: %APPDATA%\JetBrains\Junie\mcp.json

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
    }
  }
}

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"],
      "env": {
        "NOTEBOOKLM_COOKIES": "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..."
      }
    }
  }
}

Verify: Settings (Ctrl+Alt+S) → Tools → Junie → MCP Settings — notebooklm should appear in the discovered list.

Zed

~/.config/zed/settings.json — the relevant key is context_servers (Zed's MCP equivalent).

{
  "context_servers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
    }
  }
}

{
  "context_servers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"],
      "env": {
        "NOTEBOOKLM_COOKIES": "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..."
      }
    }
  }
}

Verify: Zed's AI Agent Panel lists notebooklm as an available context source.

OpenCode (sst)

opencode.jsonc (project root) or ~/.config/opencode/config.json.

{
  "mcp": {
    "servers": {
      "notebooklm": {
        "command": "npx",
        "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
      }
    }
  }
}

{
  "mcp": {
    "servers": {
      "notebooklm": {
        "command": "npx",
        "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"],
        "env": {
          "NOTEBOOKLM_COOKIES": "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..."
        }
      }
    }
  }
}

Verify: start OpenCode and ask "List my NotebookLM notebooks." — the model should call notebook_list. Gotchas: OpenCode's MCP schema is still evolving; confirm against the OpenCode docs if startup fails.

Cline (VS Code extension)

Managed via Cline's MCP settings UI inside VS Code. Underlying file (don't edit by hand): ~/.config/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json.

Open the Cline panel → click the MCP icon → Configure MCP Servers.
Paste:

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve"]
    }
  }
}

Save. Cline restarts the server automatically.

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@m4ykeldev/notebooklm-mcp", "serve", "--query-timeout", "180000"],
      "env": {
        "NOTEBOOKLM_COOKIES": "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..."
      }
    }
  }
}

Verify: ask Cline "List my NotebookLM notebooks." — the tool-call panel shows notebook_list.

Goose (Block)

~/.config/goose/config.yaml (or via goose configure). Goose calls MCP servers "extensions".

extensions:
  notebooklm:
    type: stdio
    command: npx
    args:
      - "-y"
      - "@m4ykeldev/notebooklm-mcp"
      - serve

extensions:
  notebooklm:
    type: stdio
    command: npx
    args:
      - "-y"
      - "@m4ykeldev/notebooklm-mcp"
      - serve
      - "--query-timeout"
      - "180000"
    env:
      NOTEBOOKLM_COOKIES: "SID=...; HSID=...; SSID=...; APISID=...; SAPISID=..."

Verify: goose configure → confirm the extension is enabled, then goose session and ask "List my NotebookLM notebooks."

5ire

GUI app (no user-editable config file):

Open 5ire → Settings (Mod+K → Providers).
Click Add MCP Server.
Name: notebooklm
Type: Stdio
Command: npx
Args: -y @m4ykeldev/notebooklm-mcp serve (add --query-timeout 180000 if needed)
Env vars (optional): NOTEBOOKLM_COOKIES=SID=...; HSID=...; SSID=...; APISID=...; SAPISID=...
Save and restart.

Verify: ask in chat "List my NotebookLM notebooks." — tool-call panel shows notebook_list.

Aider

Aider does not yet have first-class MCP-server client integration. Workarounds: run notebooklm-mcp serve separately and pipe responses, or use the generic stdio caller inside a small wrapper script. Track upstream support via the Aider Discord.

Generic stdio caller

If your client speaks the MCP protocol over stdio but isn't listed above, every config above reduces to the same primitive:

spawn:   npx -y @m4ykeldev/notebooklm-mcp serve
stdio:   parent <- stdout (JSON-RPC responses) | parent -> stdin (JSON-RPC requests)
env:     (optional)
  NOTEBOOKLM_COOKIES=...        # skip the auth flow in CI
  NOTEBOOKLM_CSRF_TOKEN=...
  NOTEBOOKLM_SESSION_ID=...
flags:   (optional)
  --query-timeout <ms>          # per-RPC timeout, default 120000
  --debug                       # verbose stderr logging

On connection the server advertises 32 tools — see the Complete Tool Reference below.

Verify any wiring

After restarting your client, ask:

"List my NotebookLM notebooks."

If the model invokes notebook_list and returns a table of titles, you're connected.

💬 Example Prompts

Once wired up, your AI can drive NotebookLM end-to-end with natural language. The model picks the right tool from the 32 below.

| Goal | Sample prompt | Tools the model will call | |---|---|---| | Inventory | "Show me every NotebookLM project I own" | notebook_list | | Start a project | "Create a notebook called 'Q3 Earnings' and add the AAPL 10-Q PDF at https://…" | notebook_create, notebook_add_url | | Grounded Q&A | "From the AAPL notebook, what is the year-over-year services revenue change?" | notebook_query | | Multi-source brief | "In my 'Climate Policy' notebook, generate a one-page briefing doc focused on IRA tax credits" | report_create, studio_status | | Studio podcast | "Make a 10-min audio overview of my 'AI Safety Reading' notebook" | audio_overview_create, studio_status | | Deep Research → import | "Run deep research on 'kelp aquaculture in Maine' and import the findings into my 'Climate' notebook" | research_start, research_status, research_import | | Cleanup | "Delete the audio overview I generated yesterday in 'AI Safety Reading'" | studio_status, studio_delete |

🛠 Complete Tool Reference (32)

Every tool is designed to work seamlessly within your AI's context window.

📔 Notebook Management

📄 Source Ingestion & Management

| Tool | Description | | :------------------- | :------------------------------------------------------------------------------------- | | notebook_add_url | Add any website or YouTube video as a source. Transcripts are automatically handled. | | notebook_add_text | Ingest raw text snippets or local file contents directly into your project. | | notebook_add_drive | Connect and import documents, sheets, or slides from your Google Drive. | | source_describe | Get detailed AI analysis, summaries, and key topics for any individual source. | | source_get_content | Extract the full underlying text of a source for processing by other AI tools. | | source_list_drive | List all Drive-based sources and check if they are up-to-date with the original files. | | source_sync_drive | Sync selected Google Drive sources to pull the latest changes into NotebookLM. | | source_delete | Remove a specific source from your notebook. |

🔬 Research & Deep Analysis

| Tool | Description | | :---------------- | :------------------------------------------------------------------------------------ | | research_start | Launch an autonomous research task using Google's engine (Web or Drive sources). | | research_status | Track the progress of active research tasks and view discovered insights. | | research_import | Instantly import the findings of a research task as new sources in your notebook. | | notebook_query | Ask complex, grounded questions. Answers are cited directly from your sources. | | chat_configure | Fine-tune your AI's behavior by setting specific goals or preferred response lengths. |

🎬 Studio (AI Content Generation)

| Tool | Description | | :---------------------- | :--------------------------------------------------------------------------------------- | | audio_overview_create | Transform your notebook's sources into a professional, podcast-style audio discussion. | | video_overview_create | Generate a structured video explainer based on your project data. | | report_create | Create professional Briefing Docs, Study Guides, or Blog Posts tailored to your sources. | | slide_deck_create | Turn your research into a presenter-ready slide deck automatically. | | infographic_create | Visualize complex data and relationships with an AI-generated infographic. | | flashcards_create | Generate interactive study flashcards to master your notebook's content. | | quiz_create | Create a comprehensive quiz to test knowledge grounded in your provided sources. | | data_table_create | Extract and organize information into a structured, downloadable data table. | | mind_map_create | Build a visual mind map connecting the core concepts of your notebook. | | studio_status | Check the generation status of your Studio artifacts and get download links. | | studio_delete | Clean up your workspace by deleting old Studio artifacts. |

🔑 Authentication Helpers

🎛 CLI Reference

notebooklm-mcp serve [--query-timeout <ms>] [--debug]
notebooklm-mcp auth  [--manual] [--file <path>] [--show-tokens]
notebooklm-mcp --version

| Flag | Default | Description | |---|---|---| | serve --query-timeout | 120000 | Per-RPC timeout in ms. Bump for large studio jobs or slow research. | | serve --debug | off | Verbose stderr logging of every RPC + retry. | | auth --manual | off | Skip Chrome automation; paste cookies yourself. | | auth --file <path> | — | Import a previously exported tokens JSON. | | auth --show-tokens | — | Print the cached session's metadata (cookie names, age, CSRF/SID presence — never the secret values). |

🩹 Troubleshooting

| Symptom | Likely cause | Fix | |---|---|---| | "Could not find Google Chrome or Chromium" | No Chrome/Chromium on PATH | Install Chrome or run notebooklm-mcp auth --manual | | "Authentication expired" inside the AI session | Cookies stale | The server auto-refreshes once; if it can't, re-run notebooklm-mcp auth | | Tool calls hang past 2 min | Big studio job over default timeout | Restart server with --query-timeout 300000 | | "file_path … outside the allowed roots" | notebook_add_text got a path outside cwd / tmp | Copy the file into your working directory or pass content inline | | MCP client reports server crashed on startup | dist/cli.js missing (dev clone) | pnpm install && pnpm run build | | Multiple Chrome windows pop up on concurrent failures | Older version without single-flight mutex | Upgrade to ≥ v0.2.5 |

💡 Pro Tips

Custom Timeouts: Working with massive sources? Increase the timeout: notebooklm-mcp serve --query-timeout 180000
Check Connections: Use notebooklm-mcp auth --show-tokens to verify your session validity without exposing the secrets.
CI / headless: Set NOTEBOOKLM_COOKIES (plus NOTEBOOKLM_CSRF_TOKEN, NOTEBOOKLM_SESSION_ID) to skip the browser flow entirely.

🛡 Security & Privacy

Local Storage with hardened perms. Tokens live at ~/.notebooklm-mcp/auth.json (mode 0600) inside ~/.notebooklm-mcp/ (mode 0700). Nothing leaves your machine except calls to Google.
Path-traversal guard. notebook_add_text rejects file_path arguments outside the working directory or the OS temp directory — a hostile MCP prompt cannot ask the server to read your ~/.ssh/id_rsa.
Loopback-only Chrome DevTools. The automated auth flow launches Chrome with an OS-assigned ephemeral port bound to 127.0.0.1 — no fixed-port squatting, no LAN exposure.
Single-flight auth refresh. Concurrent requests that hit an expired session share one refresh promise instead of each spawning their own Chrome.
Supply-chain gated releases. Every publish runs gitleaks → osv-scanner → pnpm audit (high+) → npm audit signatures (Sigstore) → lockfile registry pinning → publish-manifest preview before pnpm publish ships via OIDC trusted publishers. See CONTRIBUTING.md → Releasing on GitHub.
Test coverage: 100% statements / branches / functions / lines, enforced by CI.
Unofficial Tool: This project is an independent community effort and is not affiliated with Google. It interfaces with internal endpoints and may be affected by changes to the NotebookLM web platform.

📄 License

Open-source and available under the MIT License.

Crafted with precision for the AI-first developer. Part of the Model Context Protocol ecosystem.