This server, NLM Expert Research MCP — Frontier, provides authenticated, source-locked research against NotebookLM notebooks. It extracts compact evidence capsules with claims, citations, conflicts, gaps, and quality status, while managing notebooks and authentication. Key capabilities:
research — Ask a question against a specific notebook and receive an evidence capsule. Supports modes (AUTO, INSTANT, STANDARD, DEEP, VERIFY, ULTRA, FORENSIC) and internet policies (default LOCKED; also AUTO, VERIFY, AUGMENT, ULTRA).
list_notebooks — List all live NotebookLM notebooks for the authenticated account, with optional force refresh.
list_sources — Retrieve remote source IDs and titles for a notebook, useful for scoping research.
notebook_resolve — Find a notebook by title, alias, ID, or URL; triggers live refresh on cache misses.
notebook_refresh — Force refresh of notebooks and sources to detect changes (created, renamed, deleted).
get_evidence — Fetch a specific evidence item by ID from a previous session.
compare_notebooks — Explicitly compare two notebooks against a query while preserving provenance.
setup_auth — Launch a visible Chrome window for Google login, supporting 2FA and CAPTCHA.
health — Verify readiness of Frontier, authentication, and upstream transport.
doctor — Run local diagnostics for runtime, adapter, registry, and MCP operation.
certify:live — Perform certification (create/delete disposable notebooks, verify citations, soak tests, etc.) — mentioned in one version; may not be a persistently listed tool but part of the server's capabilities.
Research is source-locked by default (internet policy LOCKED), requiring explicit opt-in for external augmentation. All operations preserve provenance and are designed for deep, verifiable analysis.
Integrates with Google NotebookLM, allowing users to treat each notebook as a closed corpus, preserve evidence provenance, and retrieve compact evidence capsules. Provides tools for research, notebook resolution and refresh, evidence retrieval, and comparing notebooks.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@NLM Expert Research MCP — Frontier compare evidence on aspirin vs clopidogrel across my notebooks"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Frontier is a Windows-first Claude Desktop MCP that researches one authenticated NotebookLM notebook at a time and returns compact, source-locked evidence capsules. Its default Internet policy is LOCKED; external search requires explicit opt-in and is kept separate from canonical notebook evidence.
The published certification state is authoritative: see docs/certification.json and docs/LIVE_CERTIFICATION.md. A green local suite is not presented as live or Claude Desktop certification.
Frontier pins @roomi-fields/notebooklm-mcp exactly to 3.0.1 and applies reproducible postinstall patches for current host compatibility, citation/source provenance, RPC-only metadata, mutation safety, authentication ownership, and bounded browser lifecycle. NotebookLM has no official public API for these operations; upstream changes can still require a compatibility update.
Related MCP server: OneClickLM
npm ci
npm test
npm run setup-auth -- --force
npm run certify:live
npm run package:mcpbGoogle login, 2FA, and CAPTCHA are always completed by the user in a visible Chrome window. Normal research runs use a dedicated persistent Chrome profile and hide only Frontier-owned browser windows on Windows.
research, list_notebooks, list_sources, health, notebook_resolve, notebook_refresh, get_evidence, compare_notebooks, setup_auth, and doctor.
The upstream MCP remains an internal transport; its low-level tools are not exposed to Claude. The server writes JSON-RPC only to stdout and structured logs to stderr.
npm run certify:local
npm run check:package
npm audit --omit=dev --audit-level=high
npm run scan:secretsnpm run certify:live creates and deletes a disposable FRONTIER-CERT-* notebook, adds deterministic text, verifies citations and claim linkage over five runs, performs a ten-query soak, tests persistence and invalidation, then checks clean shutdown. It must never be substituted with mocks.
See INSTALL.md, CLAUDE_DESKTOP.md, AUTH.md, SECURITY.md, PRIVACY.md, ARCHITECTURE.md, OPERATIONS.md, and TROUBLESHOOTING.md.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Augments MCP Server - A comprehensive framework documentation provider for Claude Code
MCP server giving Claude AI access to 22+ NYC public-record databases for real estate due diligence
An MCP server for deep research or task groups
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/israelmamani/nlm-expert-research-mcp-frontier'
If you have feedback or need assistance with the MCP directory API, please join our Discord server