Skip to content
NLEN
Illustration: Gemini app now lets you create and edit files in the chat

Gemini app now lets you create and edit files in the chat

By Ivo Donker · — with AI assistance (DeepSeek V4.1 Flash)

On 29 October 2026, Google announced a feature that lets you generate, preview and edit documents, PDFs, presentations, spreadsheets, code and more in the Gemini app, directly in your chat. That is stated in a post on the Google blog The Keyword and in a post on the Google Workspace Updates blog. You no longer have to leave the chat to create a first version of a file, correct a paragraph or review a PDF.

There is, however, an important difference between platforms. You can preview and edit PDFs on both the web and mobile. All other formats, namely Google Docs, Google Sheets, slide decks, Markdown files, TXT files and code, are for now available on the web only. In this article you can read what exactly Google is announcing, how editing in the chat works, what that means for presentations, spreadsheets and documents, and which checks you can carry out yourself before you send or share something.

What Google is announcing

The announcement revolves around one move: file actions land in the chat interface of the Gemini app. According to Google, you can generate, preview and edit documents, PDFs, presentations, spreadsheets, code and more, right inside your chat. That is a different way of working from typing a prompt first, copying the answer and pasting it somewhere else. You stay in the same conversation and continue working on the file there.

Google places the announcement in two places: a blog post on The Keyword and a post on the Google Workspace Updates blog. There is also a post from @GeminiApp on X dated 29 October 2026. That post captures the essence in one sentence: "You can now generate, preview, and edit documents, PDFs, presentations, spreadsheets, and more directly in Gemini." The post also mentions the example Google itself shows in the video: turning handwritten notes into a PDF study guide, manually tweaking the content, highlighting sections so that Gemini edits text or visuals, and asking follow-up questions.

If you want to know what you can use from Google without a paid subscription, you can look at the overview of Google's free AI tools. That is useful if you want to try this feature without taking out a subscription first. How to phrase your requests to Gemini is covered in our prompt guide per model for Claude, Gemini and ChatGPT; that also helps when you steer edits to files.

What stands out about the announcement is its breadth. It is not about a single file type, but about a range of formats that you would normally open in different applications. Google itself mentions documents, PDFs, presentations, spreadsheets, code and "more", without giving a complete list of limits or exceptions in the announcement. What is explicitly set out is the platform split: PDFs on web and mobile, the rest for now web only. That split is important enough to cover separately.

How editing in the chat works

According to Google, you can expand a file to full screen in the chat and edit text directly there. You can also highlight text to ask a question, or ask Gemini to refine your files. The Workspace Updates blog describes it a little more precisely: you make manual or prompted edits inline on the web, or in a full-screen view. So there are two ways to steer things: type it yourself and apply the change directly, or ask for an adjustment in plain language.

In practice, that amounts to a way of working you probably already know from word processors, but without switching to another app. You ask for a first version, review it in the chat, expand the file when you need more room, and then highlight the part you want different. That highlighting is functional: you give Gemini a concrete spot to look at instead of a general instruction covering the whole document. That usually makes the result easier to check, because you know exactly which part was changed.

A practical habit that helps here: work in small steps. First ask for a structure or outline, then have one section worked out and assess it before you have the next one generated. That way you keep sight of what is changing and avoid having to read back a long text afterwards in which you no longer know what was original and what was generated. That is not a feature of the product, but a way of working that suits any tool that produces text for you.

If you work with documents and prompts more often, it is worth reading how to choose models for document processing. Not because this feature requires a specific model, but because the quality of what you get back is tied to how you phrase your question and which source you provide. The more concrete your instruction, the more precisely you can check it afterwards.

Presentations, spreadsheets and documents

For presentations, Google describes a clear workflow. You can create a presentation outline or slide content, preview the layout in the chat, and edit it directly or ask Gemini to adjust the tone and structure. After that you open the whole thing in Google Slides or download it as a PPTX or PDF. That means the chat is mainly the place where you create and steer the first setup, and that the further finishing happens in Google Slides. The preview in the chat is then an intermediate step, not a replacement for the editor.

For Google Sheets, the emphasis is on structure. With a prompt you can generate structured trackers, tables and data models, preview them in Gemini, and then open them directly in Google Sheets. That is useful when you quickly need a setup for, say, a list of tasks, an overview of data or a simple model, and only want to fill in the details afterwards. The preview helps you see whether the columns and rows are right before you continue in the spreadsheet itself.

For documents and PDFs, the list is a bit broader. According to Google, you can quickly draft PDFs, Google Docs, LaTeX, Markdown or TXT files, make manual or prompted edits in them, inline on the web or in full screen, and then upload them to Drive. That last part is practical: the file then stays in a place where you can find and share it, instead of existing only in a chat history.

If you first want to work out what is possible for free and where a paid subscription adds something, the article on free versus paid is a good starting point. This feature itself is not reserved for a paid tier, but the question of which extras you need elsewhere remains relevant if you work with multiple tools.

Example: from handwritten notes to a study guide

Google uses one example consistently, in the blog video and in the post on X: you upload handwritten lecture notes and ask Gemini to turn them into a PDF study guide. Then you adjust the content manually, highlight specific sections so that Gemini edits text or visuals, and ask further follow-up questions. The post on X sums that up as: "Easily turn your handwritten notes into a PDF study guide, manually tweak content, highlight sections for Gemini to edit text or visuals, and even ask follow-ups."

That example path is instructive for several reasons. First, the source is messy: handwritten notes are rarely consistent in formatting or completeness. Second, there is an intermediate step in which you do something yourself, namely adjusting things manually. Third, there is a targeted correction through highlighting, where you do not have the whole document rewritten but point to one spot. And fourth, the conversation keeps running, so you can keep asking questions about what is there.

For students, that means a workflow in which you do not first have to retype your notes before you have something usable. For knowledge workers, the same path is usable for meeting notes, a rough sketch or a collection of loose points, for example. The control, however, stays with you: a study guide that reads smoothly is not automatically a study guide that is correct. Especially with summaries of your own notes, it is sensible to check whether the emphasis has not shifted and whether anything has been added that was not in your notes.

Web and mobile: PDF only on your phone

This is the point where expectations most often diverge, so it is stated explicitly here. You can preview and edit PDFs on both the web and mobile. Google Docs, Google Sheets, slide decks, Markdown files, TXT files and code are currently supported on the web only. So on your phone you can work with a PDF, but not with those other formats.

In practice that means: if you are on the move and you want to view and adjust a PDF study guide or a PDF report, that works on mobile. If you want to create a spreadsheet setup, go through a presentation outline or edit a Markdown file, then the web version is currently the place where that is supported. So do not assume that a file type that works on your laptop gives you the same options on your phone.

That limitation is not a detail for teams that work a lot on mobile. If you are designing a workflow in which people have to update files on the move, it is sensible to plan a device with the web version for those steps, or to accept that part of the work happens at a later moment. That prevents surprises when someone in the office sees something different from what is on the phone.

Availability: all users, in all languages where the Gemini app is available

According to Google, the feature is available to all users, in all languages where the Gemini app is available. That includes Dutch-language use, in the places where the Gemini app is available.

The Workspace Updates blog names the audience explicitly: all Google Workspace customers, Workspace Individual subscribers and users with a personal Google account. The feature is not reserved for paid tiers such as Pro or Ultra. Users of all ages who have access to the Gemini app can also generate files. To get started, you ask Gemini to create the file type you want.

As for the rollout: for both Rapid Release and Scheduled Release domains, the feature is available now. If you have access to the Gemini app, you ask Gemini for the file type you need.

What this means for admins

There is no separate admin setting for this feature. Admins can turn the Gemini app on or off as a whole, following the guidance in the Help Center. That is the only control mechanism Google mentions in the announcement. So there is no switch with which you turn only the generating or editing of files on or off, separate from the app.

For admins who set policy around AI use, that means two things. First: if the Gemini app is on, this capability is in principle part of it, without an extra threshold per feature. Second: if your organisation wants specific rules about which files may and may not end up in a chat, you will have to arrange that through agreements and guidelines, not through a technical button in this feature. That makes communication more important than configuration.

It helps to have clarity in advance about which kinds of content you want to share with an AI assistant and which you do not. That is general advice, not something this announcement arranges. Think of customer data, personal data, internal assessments or material that is copyrighted. A short guideline with examples works better in practice than a long list of prohibitions.

What this means for your work

For knowledge workers, part of the work shifts to the place where you are already asking questions. A first setup of a document, a table or a presentation outline is often the part you put off longest, and that is precisely what happens here in the chat. The finishing remains human work: checking, sharpening, verifying facts and deciding what ultimately goes out. So the feature mainly changes the start, not the responsibility at the end.

For teams, it is interesting that the same environment can produce several file types. You can work on a document, a spreadsheet and a presentation outline in one conversation, and then open them in the corresponding Google apps or upload them to Drive. That makes it easier to quickly have something shareable before a meeting starts, without having to open three applications first.

For builders and developers, Google names code as one of the formats, with support on the web. That is usable for a first draft or an adjustment, but here too the same applies: check what is there before you run it or include it in a project. If you more often have to choose which model to use for which type of task, the article on choosing a model for your project helps. And if you work with long pieces of text a lot, there is the piece on models for summarising long documents, which goes into the trade-offs involved.

For students, the study guide example is the most directly applicable, especially in combination with the mobile support for PDFs. You can upload your notes, have a first version created and read and update it on your phone. The gain is in the first step; the quality of your study guide still depends on what you check and add yourself.

Checks for your own use

The announcement is fairly concrete about what is possible and where it works, and reserved about what is not yet possible. That makes it easy to get started: you ask for a file type, you review the result in the chat and you adjust what is not right. The platform split and the absence of a separate admin setting are the two points to remember before you introduce this in a team.

If you mainly use AI for study or research, the overview of AI tools for research and study gives background on other tools. That is intended as a reading tip, not as a comparison of features or prices.

Checked on 2026-10-30.

This article was compiled with AI assistance and editorially reviewed; sources are listed below.

Sources