Does Google NotebookLM have an official API, and what are the costs and free tiers associated with NotebookLM and its underlying Gemini API?

Walaikum Assalam. To clarify a common misconception: **Google NotebookLM does not currently have its own dedicated, standalone API.** NotebookLM is a consumer-facing research assistant application built on top of Google's flagship multimodal model, **Gemini 1.5 Pro**. However, depending on whether you mean the **NotebookLM web app** or the underlying **Gemini API**, here is how the pricing and free tiers work: 1. **NotebookLM Web Application (`notebooklm.google.com`):** * **Cost:** It is currently **100% free** to use via the web interface. * **Limits:** There are soft rate limits and context limits (e.g., up to 50 sources per notebook, 500,000 words per source), but no per-token charge. 2. **Underlying Gemini API (via Google AI Studio / Vertex AI):** If you want to build custom applications with the same long-context and retrieval capabilities as NotebookLM, you use the **Gemini 1.5 Pro API**: * **Free Tier:** Yes, Google AI Studio provides a generous **free tier** for Gemini 1.5 Pro (limited to 2–15 requests per minute depending on the endpoint). *Note: On the free tier, Google may use your prompts for model improvement.* * **Pay-As-You-Go Tier (Gemini 1.5 Pro):** * **Input Tokens:** $\$1.25$ per $1\text{M}$ tokens (for prompts $\le 128\text{k}$) or $\$2.50$ per $1\text{M}$ tokens (for prompts gt; 128\text{k}$). * **Output Tokens:** $\$5.00$ per $1\text{M}$ tokens (prompts $\le 128\text{k}$) or $\$10.00$ per $1\text{M}$ tokens (prompts gt; 128\text{k}$). * **Alternative (Gemini 1.5 Flash):** If you need a faster, lower-cost model for light grounding tasks, Flash costs roughly $\$0.075$ per $1\text{M}$ input tokens. If your primary goal is document-grounded Q&A via code, combining the **Gemini 1.5 Pro API** with a standard RAG pipeline (or using Google Cloud's Vertex AI Search) is the standard technical route today.