Mozilla.ai

Mozilla.ai is a small, fast-moving team building open source AI developer tools. Each of their projects lived on its own site with its own look and feel. The team needed a way to bring everything together and wanted a clear view of how people were using their docs.

For an open source developer tools team, documentation isn’t a support afterthought — it’s the front door. It’s how developers discover the tools, understand them, and decide whether to build with them.

We spoke with Anushri Gupta, who leads developer relations, documentation, and community at Mozilla.ai, about consolidating their projects into a single documentation hub — and what they’ve learned from having real analytics for the first time.

The challenge: scattered docs and no visibility

Before GitBook, each Mozilla.ai project was a documentation island — separately built, separately hosted, and visually unrelated to the others.

“Every single project had its own look and feel, and its own link. If you didn’t know any better, you wouldn’t know that Project X is built by the same team as Project Y, and that you can actually use them together.”

Compounding the fragmentation, the team’s GitHub Pages setup didn’t give them what they needed to understand their audience.

“There was no search available in our docs, and no visibility into the questions users were asking. I didn’t know what people were looking for, what users were using, what was not clear in the docs. That was the biggest driving force for me.”

A familiar Git workflow, minus the overhead

Because Mozilla.ai’s docs are closely tied to individual open source projects, keeping everything in GitHub was essential. The team already built docs on merge and checked them into a dedicated branch, and adopting GitBook meant keeping that pipeline while shedding the manual parts.

“I love the seamless integration with GitHub. It let me keep the existing pipelines, while simplifying the hard parts. I didn’t have to add any JavaScript or create HTML — just direct markdown.”

Anushri handled the migration single-handedly, adjusting scripts and GitHub Actions along the way, with engineers reviewing her pull requests. She deliberately kept the one-way sync from GitHub so the team’s existing engineering workflow stayed intact.

“I didn’t want to break the existing workflow that our engineers have. It’s easier for me to review everything in one place. They know to tag me in their PRs if there’s a documentation change, so the one-way sync just works really well for us.”

Consolidation that connects the projects

For Anushri and the Mozilla.ai team, the biggest win was bringing everything under one roof, where users could finally see how the projects related.

“Now everything is on one site. You can go from one project to another, or ask questions that give you answers across projects.”

That cross-project discovery extends to the Ask AI feature, which surfaces answers spanning multiple tools — pointing a user toward one project for a general task, then to another, like llamafile, for something more specific.

Analytics that change what gets written

Coming from GitHub Pages, the team had no analytics at all — not even basic visitor numbers. So getting a real view of their audience for the first time using GitBook’s built-in analytics has been genuinely eye-opening. And the interest goes right up to the top: the CEO likes being kept in the loop and is always glad to have fresh numbers to look at.

The analytics also surfaced how much bot and agent traffic the docs receive as the team starts to explore GEO and SEO — though the large majority of visitors are still real people reading the content.

But the most valuable signal hasn’t been who’s visiting — it’s what they’re asking. The Ask AI analytics feed directly into decisions about what to write next.

“The Ask AI feature and the analytics around it have been really helpful. We saw the same keywords appear across multiple users for llamafile. It looked like a content opportunity where we could address it directly. That’s something we wouldn’t know otherwise.”

That signal turned straight into action. The questions clustered around llamafiles and the GGUF model format, so Anushri took them to the one engineer who works on that project.

“He said, ‘Oh yeah, I don’t think the docs address this. This is something I need to write up.’ So I’m going to be working with him on writing it up and including it in our docs.”

It’s documentation insight working exactly as intended: real questions from real readers pointing straight at the gaps worth filling — a loop the team simply couldn’t see before.

Docs as the entry point

For Anushri, it all comes down to discovery. Open source software lives or dies by whether developers can find their way in.

“I’ve been in DevRel for a few years, and one thing teams miss is that your docs are the entry point to your tools or your software. If users can’t find them or don’t understand them, you’re going to lose a lot of people. For us, docs are top of the funnel — that’s how you get someone to try what you’ve built.”

Anushri has spent the past several months improving the docs, and the team is seeing it pay off — more users coming in, and Mozilla.ai starting to appear in SEO results and agentic answers in AI tools. With GitBook now hosting everything in one place, that discovery loop has a stronger foundation to build on.

For Mozilla.ai, GitBook turned a collection of disconnected project sites into a unified hub — one that looks professional, keeps the engineering workflow intact, and finally shows the team what their users actually need.

Ready to bring your documentation together in one place? Get started for free or reach out to our team to see how GitBook can simplify your documentation workflows.

Related resources:

→ Using Git Sync for collaborative documentation

→ Understand your documentation impact with insights

→ Share your product knowledge, wherever it lives, using GitBook Assistant