San Francisco-based Mintlify has raised $45 million in Series B funding at a $500 million valuation, betting that the next generation of software documentation will be read more by AI agents than by humans.
The round was co-led by Andreessen Horowitz and Salesforce Ventures, with participation from Bain Capital Ventures, Y Combinator, HubSpot Ventures, MVP Ventures, Avra, TwentyTwo Ventures, and DST Global managing partner Rahul Mehta. The deal, announced April 14, brings the company’s total funding to $67 million.
Mintlify, founded by Han Wang and Hahnbee Lee out of Y Combinator, has built a developer platform that turns codebases into navigable documentation sites with embedded AI chat. The startup now powers docs for more than 20,000 companies, with its pages reaching roughly 100 million readers a year.
But the most telling number in the announcement isn’t customer count. It’s the traffic source. Nearly half of all reads across Mintlify-hosted documentation now come from AI agents, not human developers.
Selling picks and shovels to the agent economy
The pitch is a familiar one in 2026. Build the substrate AI agents need to do real work. Coding assistants, customer-facing copilots, and autonomous research agents all hit the same wall, namely the company-specific knowledge locked inside private docs, API references, and internal wikis. If an agent can’t read it, it can’t act on it.
Mintlify’s wager is that the old documentation stack, built for humans skimming with Cmd-F, is the wrong shape for that future. Its newer features lean hard in the other direction. The platform now ships Model Context Protocol (MCP) support so external agents can query a company’s documentation directly, automated workflows that keep pages in sync as code changes, and an AI writer that drafts and updates content as engineers ship.
Andreessen Horowitz, which also led Mintlify’s Series A, sees docs turning into an API surface. Whoever owns that surface owns a piece of every AI product built on top of it.
A crowded, suddenly hot category
Documentation tooling has gone from a sleepy adjacent-to-DevRel category to one of the more contested corners of the developer tools market. Incumbents like GitBook, ReadMe, and Confluence are racing to bolt on AI features. A wave of newer entrants is trying to skip that step and reimagine the product around agents from the ground up.
That competition is showing up in churn data. Teams that adopted first-generation tools two or three years ago are now actively shopping for replacements that handle MCP, retrieval, and content automation natively. Search interest for terms like mintlify alternative and gitbook alternative has climbed sharply over the last twelve months, a sign that even the category’s biggest names aren’t running unopposed.
The pressure is mostly on pricing and openness. Mintlify’s enterprise tiers have drawn occasional grumbling on developer forums for getting expensive at scale, and some teams want more control over hosting, theming, and data residency than current SaaS doc vendors offer. Whether challengers can convert that frustration into market share remains the open question of the category.
What the money is for
The company has said the new capital will go primarily into hiring, with engineering, sales, and marketing all slated to expand. On the product side, expect more investment in the agent-facing parts of the stack. Better retrieval, deeper MCP support, and tighter loops between code changes and documentation updates. Mintlify has also been quietly extending beyond public-facing docs into internal knowledge bases, putting it on a collision course with the Notion-and-Confluence axis as much as with traditional doc tools.
The bigger story
Zoom out, and the Mintlify round fits a pattern visible across the AI infrastructure market this spring. Investors are paying premium multiples for companies that sit between proprietary data and the agents that need to reason over it. Vector databases, retrieval layers, evals, and now documentation. The thesis is that foundation models are commoditizing faster than expected, while the connective tissue between those models and enterprise data is not.
A $500 million valuation on what is still, fundamentally, a documentation product would have looked aggressive a year ago. In a world where every serious software company is shipping agents and every agent needs a knowledge source it can trust, it starts to look like table stakes. The next test is whether Mintlify can hold its lead as that pitch becomes consensus.
