minssam.
Published on

NotebookLM Is Now Gemini Notebook β€” Evolving into an Agentic Research Workstation with Code Execution

The tool that organized your notes now does the research for you.

When NotebookLM first launched, people called it "an AI that stays inside your documents." Upload a PDF and it finds answers within it β€” low hallucination, high trust. By 2026, however, that tool had become something entirely different.

Two Major Shifts

June 8, 2026 β€” The Agentic Overhaul

Google rebuilt NotebookLM on Gemini 3.5 with the Antigravity agentic framework. Three changes stand out.

β‘  A cloud computer for every notebook

Each notebook now comes with a secure cloud computer that can write and execute code. For example:

  • Upload a CSV β†’ it writes Python code automatically, runs it, and analyzes the data
  • Visualizes results as charts
  • Exports everything as a PDF report

No need to open Colab separately for data analysis anymore.

β‘‘ Automated web research β€” Deep Research

Deep Research autonomously browses hundreds of web pages and produces a structured report. It dramatically reduces the time spent on manual search β†’ read β†’ summarize cycles. Google''s own blog post describes it as being able to "conduct web research to find additional context, write code to perform accurate data analysis, and create charts and a PDF report."

β‘’ 100+ curated skills

Each notebook can access more than 100 curated skills β€” translation, summarization, code conversion, data extraction, and more. Repetitive tasks can now be automated with a single prompt.

July 16, 2026 β€” NotebookLM Becomes Gemini Notebook

The rename is more than a branding move β€” it signals full integration into the Google Gemini ecosystem. Core features like the notebook structure, source management, and Audio Overviews remain the same.

What arrived with the rename: official per-notebook code execution support.

What This Means for Real Users

Students and researchers: Upload 10 papers as sources and ask "Summarize the research trends on this topic by five-year periods." Deep Research crawls the web for additional context, and the cloud computer runs statistical analysis in code.

Content creators: Upload an industry report PDF and say "Visualize the top 10 figures as a bar chart." Gemini Notebook pulls the data, runs the code, and renders the chart.

For work: Collect meeting transcripts, email threads, and market analysis reports in one notebook. Ask the agent "Recommend three strategic options for next quarter." It reads all sources, supplements with web search, and drafts a proposal.

Things to Watch

  • Code execution and Deep Research are available to Google AI Pro subscribers and above. Some features may be limited to Google AI Ultra or Workspace enterprise users β€” verify your plan before relying on them.
  • Customizable Audio Overview length, focus, and complexity has been supported since early 2026.
  • EPUB source uploads are supported from March 2026.

Takeaway

The shift from NotebookLM to Gemini Notebook preserves the original philosophy of "staying focused on your sources" while adding web access and code execution. A simple Q&A tool has become a research workstation. Learning with AI, analyzing with AI, and finishing a report with AI β€” all in one tool.

Tips:

  • Mix source types β€” PDFs, web pages, EPUBs, and Google Docs together.
  • Be specific in Deep Research prompts; quality improves dramatically with precision.
  • Always review code execution output β€” AI-written code can still contain logical errors.

Sources:

NotebookLM Is Now Gemini Notebook β€” Evolving into an Agentic Research Workstation with Code Execution | MINSSAM.COM