At GitBook, we’re focused on building the future of documentation. Docs can no longer just be static websites with content you write by hand — they need to adapt to your users’ needs.

And when it comes to creating these adaptive docs, it’s not just about making AI write your docs for you. There are other ways to automate the creation of accurate, scalable content that adapts to how your team works.

We’ve already talked about computed content this week. And now we want to talk about what it’ll enable next: auto-translation.

We’re working on a new translation feature that lets you instantly generate translated versions of your product docs. It’s fast. It’s accurate. And it’s built on the foundations we’ve laid with computed content.

Building on computed content

You already know that GitBook’s editor is great for collaborating and publishing beautiful product documentation. But with computed content, we’re building a foundation that makes it easier to generate new documentation from other sources — rather than just writing it by hand.

On Monday, we talked about how you can use computed content to turn an OpenAPI specification into a fully structured API reference, with incredible formatting and design. Simply upload your spec file and GitBook can generate pages in seconds.

And now, we’re nearly ready to do the same for translations.

Translation tools, built in

We’re making Internationalization (i18n) a first-class part of GitBook.

And when auto-translation rolls out soon, you’ll be able to easily generate a complete copy of your chosen space in any language you want. Simply select the space you want to translate and choose the language you want to translate it into.

GitBook will analyze the content and page structure, then use AI to translate the content into your chosen language. Your new, translated content will use exactly the same layout, block structure and content as your existing space, so everything will be formatted perfectly.

It’s all built-in, so there’s no need for external tools or platforms — this is a key part of the GitBook app.

Note: This UI is still a work in progress and is not final

The power of computed content

You can translate your content as many times as you like. So if you want to localize your site into five other languages using auto-translation, simply follow the process again.

The best part? The translations all automatically update when you edit your source space.

Because of the way that computed content works, any time you make changes to your original content, GitBook will take those edits and automatically update all your translated versions as well.

That means:

• You can translate an entire docs site automatically into any language you want.

• GitBook updates every translation automatically when you edit your original content.

• You’ll get reliable, consistent translations, ready to share with users around the world.

It’s all about helping you scale your documentation — without scaling your workload.

Stay consistent with glossaries

While automated translations are powerful, we know that in product docs you don’t always want to translate every single word. Or perhaps you want to translate certain words in a specific way.

Let’s say your product uses a unique term — like “data blocks”. You don’t want that term changing across languages or showing up with odd synonyms.

For those situations, you can use glossaries.

In your glossary, you can define the terms that matter, and how you want them translated — or not translated — in each language. It helps your docs stay accurate, on-brand, and easy to follow in any language.

Note: This UI is still a work in progress and is not final

And these glossary terms will be applied across your entire docs site. So if you have multiple site sections, GitBook will translate those the titles in the navigation bar at the top of your docs site based on your glossary entries.

We’re also adding a section that lets you give the AI specific instructions to follow during the translation process. So if you want to exclude all specific block types (such as code blocks) from your translation process, for example, you can add those instructions too.

Building computed content the GitBook way

Everything we build — from computed content to improved analytics — is designed to help your team move faster.

But when we create features like computed content, we also make sure we build them in a way that lets our users leverage them and build on top of them using our API.

Right now, our integrations platform lets you create tools that plug into GitBook and add functionality or pull information from your docs to use in other tools — such as support platforms or AI code assistants. And we’re building computed content in the same way.

While we’re working on new ways to use computed content, we want you to use it in the ways that you need. In the future, you’ll be able to create integrations that pull information from existing sources of your own choosing to create computed content. Or you might want to build on our existing features, like translations, to translate using DeepL or another service.

What’s next?

OpenAPI reference docs and AI-powered auto-translation are just the beginning for computed content. We’re already thinking about more ways to use existing sources to build auto-updating content quickly and accurately.

We’re committed to making GitBook the smartest way to scale your docs — including being able to adapt that documentation to individual users’ needs, and intelligently suggest information based on what they’re working on.

We’ll have more to share on this next week.

For now, if you’re interested in being one of the first teams to test the upcoming auto-translation feature, please join the waitlist. We’re excited to hear what you think soon.

→ Get started with GitBook for free

→ Join the waitlist for auto-translation

→ Computed content: what is it and how is it helping us shape the future of docs?