Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/jacob-bd/notebooklm-mcp-cli/llms.txt

Use this file to discover all available pages before exploring further.

The research commands let NotebookLM search the web or your Google Drive on your behalf, find sources relevant to a query, and import the results directly into a notebook. This is useful for building source libraries quickly without manually hunting for articles or documents. Research tasks run asynchronously on NotebookLM’s servers. You start a task, optionally poll for its status, then import the discovered sources when it finishes — or use --auto-import to have the CLI do all three steps for you automatically. The CLI supports two command styles: 
# Noun-first
nlm research start "large language model safety" --notebook-id <id>

# Verb-first
nlm research start "large language model safety" --notebook-id <id>

​
Commands

​
research start

Kick off a research task. NotebookLM searches the web (or your Drive) and returns a set of relevant sources. 
nlm research start "query" --notebook-id <notebook-id>
Flags
FlagShortDefaultDescription
--notebook-id-nTarget notebook ID or alias (required, or use --title)
--title-tCreate a new notebook with this title and research into it
--mode-mfastResearch depth: fast (~30 s, ~10 sources) or deep (~5 min, ~40 sources, web only)
--source-swebWhere to search: web or drive
--auto-importStart, poll, and import discovered sources in one step
--force-fStart new research even if a task is already pending
--profile-pProfile to use
Use --auto-import to run the entire research workflow in a single command. The CLI will start the task, wait for it to finish (up to 2 minutes for fast mode, 10 minutes for deep mode), and immediately import all discovered sources into the notebook — no separate status or import calls needed.
# Quick web search (~30 seconds, ~10 sources)
nlm research start "transformer architecture breakthroughs 2024" \
  --notebook-id <notebook-id> \
  --mode fast

​
research status

Check the status of a running or completed research task. By default this polls until the task completes or times out. 
nlm research status <notebook-id>
nlm research status <notebook-id> --max-wait 600
Flags
FlagShortDefaultDescription
--task-id-tSpecific task ID to check (auto-detects if omitted)
--max-wait300Max seconds to poll (set to 0 for a single status check)
--poll-interval30Seconds between status checks
--compact/--full--compactShow compact or full details including source list
--profile-pProfile to use
# Single check — do not poll
nlm research status <notebook-id> --max-wait 0

# Poll for up to 10 minutes (useful for deep research)
nlm research status <notebook-id> --max-wait 600

# Show full details including discovered source titles
nlm research status <notebook-id> --full

​
research import

Import the sources discovered by a completed research task into the notebook. If TASK_ID is omitted, the CLI auto-detects the most recent completed or in-progress task. 
nlm research import <notebook-id> <task-id>
nlm research import <notebook-id>              # auto-detects the task ID
Flags
FlagShortDefaultDescription
--timeout-t300Import timeout in seconds
--cited-onlyImport only sources cited by the research report (deep mode)
--indices-iComma-separated indices of sources to import (default: all)
--profile-pProfile to use
nlm research import <notebook-id> <task-id>

​
End-to-end research workflow

1

Start research

nlm research start "EU AI Act compliance requirements" \
  --notebook-id <notebook-id> \
  --mode deep
# → Task ID: task-abc123
# → Estimated time: ~5 minutes
2

Poll until complete

# Poll with a 10-minute budget (suitable for deep mode)
nlm research status <notebook-id> --max-wait 600
# → Status: completed
# → Sources found: 38
3

Import discovered sources

# Import only the sources the report cited
nlm research import <notebook-id> task-abc123 --cited-only
# → Imported 12 sources
4

Query the newly populated notebook

nlm notebook query <notebook-id> \
  "What are the key compliance deadlines in the EU AI Act?"

​
Automated one-liner

For fast research, --auto-import collapses the entire workflow into a single command: 
nlm research start "foundation model evaluation benchmarks" \
  --notebook-id <notebook-id> \
  --mode fast \
  --auto-import
The CLI will start the task, wait for completion (up to 2 minutes for fast mode, 10 minutes for deep), and import all sources automatically.

Build docs developers (and LLMs) love

Get started for freeTalk to us