Wowerpoint

One doc in, one NotebookLM slide deck started. Slide-deck only — videos and podcasts from the same engine are noticeably worse and out of scope; refer the user to the notebooklm CLI directly if they want those.

Triggers

• "Wowerpoint <file>"
• "Make a slide deck about <file>"
• "Turn this report into slides"
• "Kawaii-deck this"

Setup (one-time per machine)

If notebooklm auth check returns 0 and command -v jq resolves, skip.

uv tool install --with playwright --force notebooklm-py
$(uv tool dir)/notebooklm-py/bin/playwright install chromium

jq is required by the workflow's JSON parsing; install if missing (brew install jq on macOS, or your distro's package manager).

Then the user authenticates interactively — do not script. Tell them to type ! notebooklm login so the OAuth ENTER lands in their terminal.

Workflow

1. The source doc

You need exactly one source doc. If it doesn't exist or is too thin to carry a deck, write it first — use mem-search and sequential thinking to make it comprehensive (long-form, narrative, several thousand words is normal). Do not paper over a weak source by adding more sources.

2. Auth pre-flight

notebooklm auth check 2>&1 | tail -5

Exit 1 with Run 'notebooklm login' to authenticate. = halt and tell the user.

3. Create the notebook

NOTEBOOK_ID=$(notebooklm create "<title>" --json | jq -r .notebook.id)

Title: H1 of the source doc, or its filename stem; append a date for dated work.

4. Add the source

SOURCE_ID=$(notebooklm source add "<doc-path>" --notebook "$NOTEBOOK_ID" --json | jq -r .source.id)

The JSON envelope keys are create → .notebook.id and source add → .source.id. A wrong key produces an empty string and a silent downstream failure.

5. Wait for the source to be ready

notebooklm source wait <SOURCE_ID> -n <NOTEBOOK_ID> --timeout 600

6. Start generation

notebooklm generate slide-deck "<PROMPT>" --format detailed --length default --notebook <NOTEBOOK_ID> --json --retry 2

7. Print the notebook URL and end the turn

After starting generation, print:

https://notebooklm.google.com/notebook/<NOTEBOOK_ID>

Then end the turn.

Output path

Adjacent to the source, parallel filename:

<source-dir>/<source-stem>-slides.pdf

If the source isn't somewhere that makes sense as an output location, default to reports/<stem>-slides.pdf.

PDF on explicit request

Only if the user later explicitly asks for the PDF, use the output path above:

notebooklm artifact list -n <NOTEBOOK_ID>
notebooklm artifact wait <artifact_id> -n <NOTEBOOK_ID> --timeout 1800
notebooklm download slide-deck "<OUTPUT_PATH>" -a <artifact_id> -n <NOTEBOOK_ID>

The deck may still be generating, so always wait on the artifact before downloading.

The prompt

One sentence. Default:

Use kawaii characters to tell the story of <subject>. Keep it warm and clear.

Replace <subject> with a one-phrase description from the source doc's H1 or the user's framing. If the user supplies their own prompt, pass it through verbatim — don't expand it.

Failure modes

• pip: command not found — modern macOS doesn't ship pip on PATH. Use uv tool install.
• Playwright not installed — install notebooklm-py with --with playwright, then playwright install chromium.
• Run 'notebooklm login' to authenticate — only the user can complete OAuth.
• Rate-limit (GENERATION_FAILED or "No result found for RPC ID") — --retry 2 handles transients. If it still fails, report the error and end the turn; retry or use the web UI only on a later user request.
• Source upload denied for sensitive docs — confirm before adding sources containing credentials, customer data, or unreleased product info. NotebookLM is a Google service.
• --length long does not exist — only default|short. If the user asks for "long slides," use default and explain.
• No --style flag — kawaii lives in the prompt text.

Operational tips

• Rerun cheaply — once the notebook + source exist, regenerating with a different prompt only repeats generation. Reuse NOTEBOOK_ID and SOURCE_ID.
• Web UI fallback — if the user later reports generation is rate-limited for more than 30 minutes, open the notebook URL and trigger generation in the UI. Download a PDF only if they explicitly request it.