agentleFS
julianoczkowskijulianoczkowski/notebooklm-mcp-2026/CLAUDE.md

notebooklm-mcp-2026

CLAUDE.md16 starsChanged 9 months agoMITView on GitHub

An MCP (Model Context Protocol) server that lets AI assistants query Google NotebookLM notebooks. Authenticates via Chrome CDP cookie extraction, communicates with Google's internal batchexecute RPC API, and exposes 9 tools over stdio JSON-RPC. Built with FastMCP, httpx, and Rich. All CLI subcommands support --debug for verbose logging. The setup command also supports --dry-run. 5-layer stack, each layer only calls the one below it: Data flow: MCP JSON-RPC → server.py dispatches → tool function → get_client() returns singleton → NotebookLMClient…

  • Installs packages

The file

We split this CLAUDE.md into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    All CLI subcommands support --debug for verbose logging.

    line 35
  2. Very specific:

    All tool functions return {"status": "success"|"error", ...} dicts

    line 56
  3. Very specific:

    All constants live in config.py — no other module hardcodes URLs, RPC IDs, headers, or timeouts

    line 57

≈ “Common” is an estimate: at least 1 in 5 of the 28,029 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

What This Project Is

An MCP (Model Context Protocol) server that lets AI assistants query Google NotebookLM notebooks. Authenticates via Chrome CDP cookie extraction, communicates with Google's internal batchexecute RPC API, and exposes 9 tools over stdio JSON-RPC. Built with FastMCP, httpx, and Rich.

Commands

# Install
pip install -e ".[dev]"

# Lint
ruff check src/ tests/

# Test
pytest -v --tb=short
pytest -k "not chrome"              # skip Chrome-dependent tests
pytest tests/test_protocol.py -v    # run one test file
pytest -k "test_build_rpc_body"     # run one test by name

# Run
notebooklm-mcp-2026 serve           # start MCP server (stdio)
notebooklm-mcp-2026 login           # interactive Chrome login
notebooklm-mcp-2026 logout          # remove stored credentials
notebooklm-mcp-2026 setup           # interactive setup wizard
notebooklm-mcp-2026 doctor          # diagnose common issues
notebooklm-mcp-2026 status          # show auth + config status
notebooklm-mcp-2026 version         # print version
notebooklm-mcp-2026 help            # show help message

All CLI subcommands support --debug for verbose logging. The setup command also supports --dry-run.

Architecture

5-layer stack, each layer only calls the one below it:

MCP Clients (Claude Desktop, etc.)
  └─ server.py        — FastMCP instance + global client singleton (get_client())
       └─ tools/      — 9 tool functions: validate input → call client → return dict
            └─ client.py   — NotebookLMClient: HTTP requests, auth retry, backoff
                 └─ protocol.py  — Pure functions: build RPC bodies, parse responses (zero I/O)
                    auth.py      — Chrome CDP cookie extraction + secure disk storage

Data flow: MCP JSON-RPC → server.py dispatches → tool function → get_client() returns singleton → NotebookLMClient method → protocol.build_*() encodes request → httpx POST → protocol.parse_*() decodes → dict response back up the stack.

Key patterns:

  • server.get_client() lazy-initializes a singleton NotebookLMClient from disk-cached tokens
  • protocol.py has zero I/O — all encode/decode functions are pure and directly unit-testable
  • 2-layer auth recovery: CSRF refresh first, then full token reload from disk
  • All tool functions return {"status": "success"|"error", ...} dicts
  • All constants live in config.py — no other module hardcodes URLs, RPC IDs, headers, or timeouts
  • CLI output goes to stderr (stdout reserved for MCP stdio transport)

Testing

  • conftest.py provides fixtures: sample_cookies, sample_batchexecute_response, sample_query_response
  • test_tools.py — unit tests for all 9 MCP tool functions (mock server.get_client and auth.load_tokens)
  • @pytest.mark.chrome — tests requiring Chrome, auto-skipped in CI
  • @pytest.mark.integration — integration tests using FastMCP Client
  • asyncio_mode = "auto" — async tests just work without explicit markers
  • protocol.py tests need no mocking (pure functions)
  • Coverage: pytest --cov=notebooklm_mcp_2026 --cov-report=term-missing

Code Conventions

  • Python 3.11+, from __future__ import annotations everywhere
  • Ruff: line-length 100, target py311
  • Google-style docstrings (Args, Returns, Raises)
  • Custom exceptions in client.py: NotebookJulianError (base), AuthenticationError, APIError, ValidationError
  • Credentials stored at platformdirs.user_data_dir("notebooklm-mcp-2026")/auth.json with 0o600 permissions
  • Pre-commit hooks: ruff check + ruff format (.pre-commit-config.yaml)

Environment Variables

  • NOTEBOOKLM_MCP_DATA_DIR — override storage directory
  • NOTEBOOKLM_BL — override Google build label when it rotates
  • NOTEBOOKLM_QUERY_TIMEOUT — override query timeout (default 120s)

CI

  • GitHub Actions: runs on ubuntu, macos, windows × Python 3.11, 3.12, 3.13
  • Ruff lint + pytest with coverage on every push/PR
  • PyPI publish triggered by v* tags via trusted publishing

Repo Structure

  • .github/FUNDING.yml — sponsor links (Buy Me a Coffee, Ko-fi)
  • .github/CODEOWNERS — @julianoczkowski as default reviewer
  • .github/dependabot.yml — weekly dependency updates (pip + github-actions)
  • CONTRIBUTING.md — contributor guide
  • CODE_OF_CONDUCT.md — Contributor Covenant v2.1
  • CHANGELOG.md — release history
  • SECURITY.md — security policy + threat model
  • .github/ISSUE_TEMPLATE/ — bug report and feature request templates
  • examples/ — usage examples (basic_workflow.py, follow_up_conversation.py)

Skills (Slash Commands)

  • /release — lint, test, version bump, git tag, GitHub release, PyPI publish via CI
  • /cleanup-branches — delete merged local + remote branches, prune stale refs
  • /sync-docs — scan recent commits, update CLAUDE.md and CHANGELOG.md to match

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.