According to Android Authority, Google has finally made chat history available to all users of its AI research service, NotebookLM, on both mobile and the web. This key feature, which allows for persistent chat sessions across different devices, was first announced back in October with a promised rollout starting the following week. The update means you can now start a research conversation on your phone and seamlessly continue it on your laptop. This joins two other recent additions: integration with the main Gemini chatbot earlier this month and a massive boost to custom chat prompts, from 500 characters to 10,000. The team confirmed the rollout on the NotebookLM Twitter account, noting that you can delete history anytime and, in shared notebooks, your chats remain private to you.
Why this took so long
Look, let’s be honest. A basic feature like chat history shouldn’t be a six-month rollout saga. Google first mentioned this in October, and here we are. It makes you wonder about the backend complexity for a product that’s essentially a fancy wrapper around your own documents and a large language model. Was it a sync issue? A privacy overhaul? Who knows. But the delay highlights a classic problem with Google’s experimental apps: they often launch in a surprisingly bare-bones state. The core idea of NotebookLM—turning your sources into a conversational partner—is brilliant. But without something as fundamental as remembering your last question, its utility was hamstrung. Basically, it wasn’t a proper tool you could rely on for ongoing work.
The real impact on users
So why does this actually matter? It changes NotebookLM from a neat demo into a usable research companion. Think about it. You’re deep in a complex analysis of a set of PDFs for a project. You ask a few questions on your desktop, get some great leads, and then have to run out. Before, that thread was gone. Poof. Now, you can pull out your phone on the train and pick up right where you left off. That’s a game-changer for continuity. The boost to custom prompts—from 500 to 10,000 characters—is arguably just as big. It means you can feed the AI entire paragraphs of context, detailed instructions, or multi-step reasoning frameworks. You’re not just asking questions anymore; you’re programming a specialized research agent. These aren’t flashy features. They’re the boring, essential plumbing that makes an AI tool stick.
Where NotebookLM fits now
With Gemini integration and these foundational upgrades, NotebookLM is starting to carve out a clearer niche. It’s not trying to be your everything chatbot like ChatGPT or Gemini. Instead, it’s positioning itself as the AI for focused, source-grounded work. The announcement post even hints at the next frontier: “Now, about adding support for drawings, flowcharts, and graphs.” That’s the signal. They’re aiming for the academic, analyst, and content creator crowd—people who think in visual diagrams and need to synthesize information from proprietary documents. Is it there yet? Not quite. But with persistent chat and massive prompts, the foundation is finally solid. The wait was frustrating, but the tool is now significantly more legitimate.
