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.
Notebooks are the top-level containers in NotebookLM — everything else (sources, studio artifacts, research tasks) lives inside one. The nlm notebook commands let you manage the full notebook lifecycle from creation through deletion, including an AI-powered describe command that summarises what is inside a notebook and suggests follow-up topics.
The CLI supports two command styles that are fully equivalent — use whichever feels more natural in your workflow:
# Noun-first (resource-oriented)
nlm notebook create "My Research"
# Verb-first (action-oriented)
nlm create notebook "My Research"
Commands
notebook list
List all notebooks in your account.
nlm notebook list
nlm list notebooks # verb-first alias
| Flag | Short | Description |
|---|
--json | -j | Output as JSON |
--quiet | -q | Output IDs only (one per line) |
--full | -a | Show all columns |
--title | -t | Output as ID: Title format |
--profile | -p | Profile to use |
| Flag | Output style | Best for |
|---|
| (none) | Rich table with key columns | Human browsing |
--json | Full JSON array | Scripting / parsing |
--quiet | One ID per line | Shell piping |
--full | Table with all columns | Detailed inspection |
--title | ID: Title per line | Quick reference |
notebook create
Create a new, empty notebook.
nlm notebook create "AI Research 2025"
nlm create notebook "AI Research 2025" # verb-first
--quiet or capture with --json for scripting.
Flags
| Flag | Short | Description |
|---|
--json | -j | Output result as JSON |
--profile | -p | Profile to use |
notebook get
Retrieve full details for a single notebook by ID or alias.
nlm notebook get <notebook-id>
nlm notebook get myproject # alias resolves automatically
| Flag | Short | Description |
|---|
--json | -j | Output as JSON |
--profile | -p | Profile to use |
notebook describe
Ask the AI to summarise the notebook and suggest relevant topics to explore. This calls the NotebookLM AI layer — not a local operation — so it requires a valid session. The response includes an AI-generated summary of all sources and a list of suggested follow-up questions.
nlm notebook describe <notebook-id>
| Flag | Short | Description |
|---|
--json | -j | Output as JSON |
--profile | -p | Profile to use |
nlm notebook describe produces an AI-generated summary of the notebook’s sources and suggested topics — it does not just return raw metadata. nlm notebook get returns the raw notebook metadata (title, ID, source count) without AI processing.
notebook rename
Rename a notebook in place.
nlm notebook rename <notebook-id> "New Title"
| Flag | Short | Description |
|---|
--profile | -p | Profile to use |
notebook delete
Permanently delete a notebook and all its sources and artifacts.
This action is irreversible. All sources, studio artifacts, and research tasks inside the notebook are destroyed. Always use --confirm in scripts to skip the interactive prompt.
nlm notebook delete <notebook-id> --confirm
--confirm the CLI will prompt you interactively before proceeding.
Flags
| Flag | Short | Description |
|---|
--confirm | -y | Skip the confirmation prompt |
--profile | -p | Profile to use |
notebook query
Chat with the sources inside a notebook. This is a single-turn query; for multi-turn conversations keep the conversation ID between calls.
nlm notebook query <notebook-id> "What are the main arguments?"
nlm notebook query <notebook-id> "Follow up question" --conversation-id <conv-id>
| Flag | Short | Description |
|---|
--json | -j | Output response as JSON |
--conversation-id | -c | Conversation ID for follow-up questions |
--source-ids | -s | Comma-separated source IDs to restrict the query |
--timeout | -t | Query timeout in seconds (default 120) |
--profile | -p | Profile to use |
Scripting examples
Shell — pipe IDs with --quiet
Use --quiet when you only need IDs — it outputs one ID per line with no extra formatting, making it ideal for shell piping.
# Get the first notebook ID and query it
notebook_id=$(nlm notebook list --quiet | head -1)
nlm notebook query "$notebook_id" "Summarize the key points"
# Delete all notebooks whose titles start with "Draft" (careful!)
nlm notebook list --quiet | while read -r id; do
nlm notebook delete "$id" --confirm
done
Python — parse --json output
When you need metadata beyond just the ID (title, source count, creation time), use --json and parse with the standard library.
import json
import subprocess
result = subprocess.run(
["nlm", "notebook", "list", "--json"],
capture_output=True,
text=True,
check=True,
)
notebooks = json.loads(result.stdout)
for nb in notebooks:
notebook_id = nb["id"]
title = nb.get("title", "(untitled)")
print(f"{notebook_id}: {title}")
# Query the first notebook
first_id = notebooks[0]["id"]
query_result = subprocess.run(
["nlm", "notebook", "query", first_id, "Summarize the key points", "--json"],
capture_output=True,
text=True,
check=True,
)
response = json.loads(query_result.stdout)
print(response["answer"])
Complete notebook workflow
Create a notebook
nlm notebook create "Q3 Competitive Analysis"
# → Notebook ID: abc12345-...
Set an alias for convenience
nlm alias set q3 abc12345-...
Add sources
nlm source add q3 --url "https://example.com/report" --wait
nlm source add q3 --file analysis.pdf --wait
Query the notebook
nlm notebook query q3 "What are the top three competitor differentiators?"