notebooklm-mcp

Provides an MCP server that lets local AI coding agents (e.g., Claude Code, Cursor, Codex) interact with Google NotebookLM to answer questions grounded in user-owned NotebookLM notebooks, including session/library management and browser-automation-based authentication.

Evaluated Mar 30, 2026 (21d ago)
Repo ↗ Ai Ml mcp notebooklm gemini browser-automation typescript documentation research agent-tools coding-assistance
⚙ Agent Friendliness
64
/ 100
Can an agent use this?
🔒 Security
63
/ 100
Is it safe for agents?
⚡ Reliability
35
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
78
Documentation
65
Error Messages
0
Auth Simplicity
60
Rate Limits
50

🔒 Security

TLS Enforcement
95
Auth Strength
55
Scope Granularity
40
Dep. Hygiene
60
Secret Handling
65

Strengths inferred from README: authentication happens locally via a Chrome window and the README claims credentials never leave the machine. Concerns: it performs browser automation against a third-party service; handling of persisted sessions/tokens and their storage/logging behavior is not evidenced in the provided content. Rate limits and security controls are not specified (no scope model, no mention of fine-grained access control). Dependency list is small but no CVE/security posture is provided.

⚡ Reliability

Uptime/SLA
0
Version Stability
55
Breaking Changes
50
Error Recovery
35
AF Security Reliability

Best When

You want citation-backed answers from NotebookLM for internal docs/notebooks and your workflow uses MCP-capable coding agents.

Avoid When

You cannot use browser automation, need strict auditability of automation actions, or require guaranteed continuity despite upstream NotebookLM/Gemini UI or access changes.

Use Cases

  • Ask grounded questions about your own documents/notebooks while coding
  • Research implementation details and edge cases from NotebookLM before generating code
  • Maintain and select a NotebookLM “library” (tagged links) shared across multiple MCP-capable clients
  • Run multi-turn research sessions that progressively gather details from NotebookLM
  • Clean up stored NotebookLM automation data via provided MCP tools

Not For

  • Replacing or bypassing NotebookLM for sensitive data you cannot share with Google services
  • Guaranteeing compliance/contractual data-residency requirements for enterprise environments
  • Production-grade automation where browser-automation reliability and ToS impact are unacceptable

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

Methods: Browser-based login to NotebookLM via Google account (one-time flow triggered by MCP tool) Session persistence and re-authentication tools (per tool profiles/commands described)
OAuth: No Scopes: No

Authentication is initiated by prompting the user (e.g., “Log me in to NotebookLM”), which opens a Chrome window for Google login. The README claims credentials remain on the machine, but no concrete technical details (e.g., token storage format) are provided in the included content.

Pricing

Free tier: No
Requires CC: No

The server itself is distributed via npm (MIT) and pricing is not described beyond NotebookLM/Google account limits.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • Browser automation can be sensitive to upstream NotebookLM UI changes and authentication/consent flows
  • Rate limits are governed by NotebookLM/Google account query limits; retries may worsen throttling
  • Tool context load varies by profile; using too many tools can increase latency/cost
  • Session/library state may persist; agents should use explicit select/reset/cleanup commands to avoid stale context

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for notebooklm-mcp.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

8642
Packages Evaluated
17761
Need Evaluation
586
Need Re-evaluation
Community Powered