What problem does it solve? Google NotebookLM has no public API, so integrating it into an app requires browser automation and session management. This Skill explains how the notebooklm-api-client package bridges that gap with a Cloudflare Worker, a Puppeteer-driven Google login, and a container running notebooklm-py behind a single authenticated POST endpoint. ## Core Features & Use Cases - Single POST API: Route six actions (login, list, create, ask, summarize, delete) through one endpoint secured by a Bearer API_TOKEN. - Automated Google Login: Drive Puppeteer through the Google sign-in flow, handle 2FA security codes, and persist cookies and localStorage to the container's auth file. - Container Lifecycle Management: Understand the NotebookRunner Durable Object container that sleeps after 5 minutes idle and shares one auth state across all traffic. - Use Case: You want your app to ask questions against a NotebookLM notebook. Deploy the Worker with wrangler, POST a login action to capture the Google session, then POST ask actions with a notebookId and prompt to get answers. ## Quick Start Ask the assistant to deploy the notebooklm-api-client Worker with wrangler and send a login action followed by an ask action against a notebook.
