Gemini Notebook
Google's grounded research notebook, renamed from NotebookLM in July 2026. Upload sources, get answers with inline citations and run code against them in a per-notebook sandbox.
Overview
The product works on one constraint: it answers from sources you give it. Upload PDFs, Google Docs, web pages, YouTube transcripts or pasted text, and every response comes back with inline citations. That constraint is the whole value. For an engineer working through forty papers on visual SLAM or a vendor's 300-page integration spec, the gap between a grounded answer and a merely plausible one is the entire job.
July 2026 brought two changes. Google renamed it from NotebookLM to Gemini Notebook, and gave each notebook a sandboxed cloud computer so it can write and execute code against your uploaded material rather than only reading it. Audio and video overviews turn a source set into a narrated walkthrough, which is more useful for briefing a teammate than for your own study. Free accounts get 50 sources per notebook. Google AI Pro at $19.99 a month raises that to 300, and Ultra tiers start at $99.99. Google puts usage at more than 30 million people and 600,000 organizations.
Teams with documents they cannot send to Google should skip it. There is no on-prem option, no bring-your-own-key and no self-hosting path, so anything under NDA or export control stays out. Ingestion also struggles with two-column academic PDFs full of equations, where formulas and figures get flattened. Test it against your actual corpus before you build a workflow on top of it.
Key Features
- ✓ Answers grounded strictly in your uploaded sources, with inline citations pointing at the exact passage
- ✓ Native code execution inside a secure per-notebook cloud computer, shipped July 2026, for analysis over your own material
- ✓ 50 sources per notebook on free, 300 on the Google AI Pro tier, 500 or more on Ultra
- ✓ Audio and video overviews that turn a source set into a narrated walkthrough for handing context to someone else
- ✓ Deep Research mode that goes out and gathers sources rather than only reading what you upload
- ✓ Notebooks now sync across the Gemini app and Google Search
Where it holds
- • Citation grounding is the strongest of any mainstream research tool: every claim points at a passage you can go check yourself
- • The free tier is genuinely usable, and 50 sources per notebook covers most real technical reading projects
- • Code execution over your own uploaded data closes much of the gap with spinning up a scratch Jupyter notebook
Where it breaks
- • The July 2026 rename broke bookmarks, internal docs and links across teams, and the rollout was staggered over several weeks
- • New capability lands on AI Ultra and Workspace Business first, so $19.99 Pro subscribers wait
- • No on-prem build, no bring-your-own-key and no self-hosting, which rules it out for NDA, export-controlled or air-gapped material
- • Ingestion still flattens dense two-column academic PDFs, and equations and figures come through as noise
My Take
The July 2026 rename from NotebookLM arrived alongside the part that actually matters: every notebook now gets a secure cloud computer that can write and run code against your own sources. That shifts it from a reading tool toward a scratch analysis environment. Citation grounding is still the reason to pick it over a general chat window, because each answer points at the passage it came from instead of asking you to trust it. The catch is release order, with new capability going to AI Ultra and Workspace Business customers first while Pro subscribers at $19.99 wait weeks.
Quick Info
- Pricing:
- freemium
- Openness:
- Proprietary
- Starting at:
- Free tier gives 50 sources per notebook plus a daily chat cap. Paid access rides on Google AI subscriptions: AI Plus $4.99/month, AI Pro $19.99/month (300 sources per notebook), AI Ultra from $99.99/month. Google cut the Plus tier from $7.99 to $4.99 in June 2026.
- Added:
- Aug 2026
- Updated:
- Aug 2026
Use Cases
Judge it on your own work
The notes above say where Gemini Notebook holds and where it breaks. The fastest check is your own workload.
Visit website ↗Alternatives to Gemini Notebook
Elicit
freemiumLiterature search and systematic review across 138 million papers, with screening and extraction accuracy figures the vendor actually publishes.
Glean
paidEnterprise search and agent platform that indexes company systems with permissions intact, then exposes that context to internal agents and outside AI clients.
RAGFlow
open-sourceInfiniFlow's Apache 2.0 RAG engine built on deep document parsing. DeepDoc handles layout, tables and OCR before chunking, and every citation points back to a highlighted region in the source PDF.