2025 was the most transformative year in Mintlify's history. What started as a documentation platform evolved into something much larger: the infrastructure layer for how AI understands and interacts with technical knowledge.

Here's what we accomplished together.

• Scaled from low 7-figures to over 8-figures in ARR
• Grew from 12 people to 40, including our first acquisition with Trieve
• Onboarded over 10,000 companies including Coinbase, Microsoft, AT&T, PayPal, Perplexity, and many more
• Processed over 1 million documentation updates
• Powered content viewed over 280 million times by people worldwide
• AI usage grew 8x year-over-year
• Moved offices… twice

For the past decade, documentation has been static. You write it, you publish it, and then you hope people can find what they need. That model is breaking down.

AI is fundamentally changing how users discover and consume technical information. They are not browsing sidebar navigation anymore. They're asking questions and expecting answers. This shift isn't coming. It's here. And it demands an entirely new approach to documentation infrastructure.

This year, we rebuilt Mintlify from the ground up to meet this moment.

For Users Reading Content

We're now serving over one million AI assistant queries every month directly within the documentation experience.

This isn't a chatbot bolted onto the side of your docs. The assistant is built into the product itself, understanding your content deeply enough to answer questions with accuracy and context. When developers ask a question, they get real answers drawn from your documentation, not hallucinated responses or generic suggestions.

Documentation is no longer just something you read. It's something you talk to.

For AI Systems Ingesting Content

The way AI agents consume documentation matters enormously. Most docs weren't built for machines to read—they were built for humans to browse. That's a problem when the majority of your "readers" are increasingly LLMs, coding assistants, and autonomous agents.

We've been working to change that:

Markdown responses: Documentation now returns clean, structured markdown that AI systems can actually parse and use Content negotiation: Serve different formats based on who's asking - human or machine Pioneering llms.txt: We helped establish the standard for how AI should discover and interact with documentation MCPs as part of docs: Model Context Protocol integration means your documentation becomes a tool that AI agents can directly invoke One-click LLM copy: Making it trivial to get your docs into any AI context window

We're building the rails for how AI understands technical products. When an agent needs to know how your API works, we make sure it gets the right answer.

For Teams Writing Content

Writing documentation has always been a tax on engineering velocity. You ship a feature, then you have to stop and document it. Then the feature changes, and the docs fall out of sync. It's a losing battle.

This year, we launched something that changes the equation: the Mintlify Agent.

The agent doesn't just help you write documentation. It keeps documentation alive. It monitors your codebase, detects changes, and proposes updates automatically. Self-updating docs aren't a dream anymore. They're shipping to customers today.

Documentation is no longer something you maintain. It's something that maintains itself.

Behind the vision, there's a relentless pace of execution. Our team merged over 6,000 pull requests this year.

Some highlights:

Performance: Every doc page now loads instantly. We obsessed over milliseconds because developer experience compounds. New themes: 8 new themes launched, with more coming. Your docs should look as good as your product. New editing experience: A complete overhaul of how you create and manage content. New API Playground: An entirely rebuilt experience for interactive API exploration.

And countless smaller improvements that would take too long to list.

In total, we shipped over 6000 pull requests.

We started Mintlify because documentation was broken—outdated the moment it was written, impossible to maintain, and increasingly irrelevant in how developers actually work.

This year proved that there's a better way. Documentation can be intelligent. It can understand questions and provide answers. It can serve both humans and AI systems. It can update itself.

2026 will be about building towards the future of knowledge management: self-updating documentation and the infrastructure layer for knowledge.

To every company trusting us with their documentation, to every team member who shipped through the chaos, and to everyone who believes developer experience matters: thank you.

Here's to building what comes next.

Han Wang CEO of Mintlify