From scribbles to searchable ideas: how Google’s InkSight matured

Bridging the gap between your physical notebook and digital workflow with refined AI intelligence.

My 2024 InkSight Article Revisited

I wrote in November 2024 about Google’s InkSight. Time to look what has happened to this initiative since then. Back then, I was fascinated by the idea of technology that could bridge the chasm between our tangible, scribbled notes and the sleek, searchable world of digital text. For a writer who often finds inspiration in the messy act of putting pen to paper, brainstorming or outlining my articles, InkSight felt like a glimpse into a future where my physical notebooks could finally talk to my digital tools.

In short: Google’s InkSight has evolved from a research experiment into a foundational AI capability integrated within the Gemini 3 ecosystem. Rather than a standalone app, its “derendering” technology now powers improved handwriting recognition in NotebookLM and Google Keep. This shift allows writers to seamlessly transition from physical brainstorming to digital drafts with higher accuracy.

image created by me using Nano Banana in Google Gemini

My article, “From Scribbles to Searchable Text,” explored InkSight’s promise: transforming photos of handwriting into editable, digital ink. The potential was clear: imagine photographing your latest stream-of-consciousness brainstorm and having it instantly ready for refinement in Google Docs, or easily searchable within Google Keep. This wasn’t just about OCR; it was about understanding the flow of our thoughts, captured in ink.

So, what’s happened to InkSight since then? Has it launched as a standalone app, a magical “Ink to Pixel” button? The answer, as is often the case with powerful underlying technologies, is both more subtle and more impactful: InkSight hasn’t become a product; it’s become the intelligence woven into the products we already use.

The Silent Evolution: InkSight Inside

The core innovation of InkSight, the “derendering” technology that reconstructs handwriting from an image into vector-based strokes, has matured significantly. In June 2025, the research was formally published in the Transactions on Machine Learning Research (TMLR). This academic validation confirmed its robustness and moved it from a research experiment to a foundational AI capability.

See: InkSight: Offline-to-Online Handwriting Conversion by Teaching Vision-Language Models to Read and Write

But for us, the creatives and the note-takers, the real magic is in its integration:

1. Gemini’s New Eyes: Reading Your Real-World Notes

Google’s Gemini models, especially the latest Gemini 3 Pro and Flash (released in late 2025), are now infused with the principles behind InkSight. This is why tools like NotebookLM, which many of us rely on for digesting research and generating insights, have become incredibly adept at “reading” handwriting.

Think about it: before, uploading a PDF of your handwritten meeting notes or a photo of your brainstorming whiteboard often resulted in a jumbled mess of OCR text. Now, Gemini, powered by InkSight’s deep understanding of visual context, can not only extract the words but also grasp the layout, connections, and intent of your scribbles. It’s no longer just recognizing characters; it’s interpreting your handwritten thoughts.

2. Your Keep Notes, Now Smarter Than Ever

For those of us living in Google Keep, InkSight’s quiet influence is making itself felt. While we might not have a “convert to digital ink” button (yet!), the backend intelligence allows Gemini to interact with your Keep notes in new ways. Have a photo of a handwritten recipe or a quick thought you jotted down? You can now prompt Gemini to “Find the note about the cinnamon muffins” or “Summarize the key ideas from my brainstorm on the article about AI ethics.” The AI is seeing through the ink to the ideas.

3. NotebookLM as Your Agentic Muse

This is perhaps the most exciting development for us as writers. With the “Agentic” features introduced to NotebookLM in late 2025, you can upload those raw, handwritten drafts — the ones full of arrows, crossed-out words, and circled ideas — and have the AI act as a true creative partner. It can:

• Expand on a circled idea: “Take this point about ‘narrative voice’ and give me three examples.”
• Outline a section: “Based on these bullet points, draft a quick outline for the intro to my article on overcoming writer’s block.”
• Suggest connections: “I wrote this paragraph on AI developments here, and this one on content creation there. Are there any natural links between them?”

This isn’t just digitizing; it’s collaborating with your own handwritten thoughts, letting AI bridge the gap between your spontaneous bursts of creativity and structured output.

The Future is Fluid (Ink)

My 2024 article concluded by saying the wall between physical and digital was crumbling. Today, I can confidently say that wall is largely gone. InkSight, while operating behind the scenes, has fundamentally changed how Google’s AI interacts with our handwritten world.

For “The Carefree Writer,” this means even more freedom. Brainstorm anywhere, anytime, with your favorite pen and paper. The digital world is now ready to meet you there, ready to interpret, refine, and collaborate with your most authentic, ink-stained ideas.

What’s next? Perhaps direct “beautification” of handwritten notes in Keep, transforming your hasty scribbles into elegant, editable text at a tap. Or even more intuitive “AI agents” that learn your unique handwriting style and anticipate your needs. The pen may be mightier than the sword, but when paired with AI insights, it’s becoming an unstoppable force for creativity.

SEO Description:

Keywords / Tags:

• Writing Skills
• Google Gemini
• NotebookLM
• Digital Productivity
• Handwriting Recognition