The Mintlify agent just got significantly better. It can now handle more complex documentation tasks like processing files and images, making updates across multiple pages, and addressing comments in pull requests.

As the solo writer leading docs at Mintlify, I love that the agent helps more people across the company create high quality documentation PRs and that it reduces the number of tasks stuck in the backlog.

There are a few engineers at Mintlify who write the docs for every feature that they ship. They're great writers and enjoy teaching people how to get the most from what they build.

I don't expect everyone to contribute to the docs, but I want everyone to be able to. Even if they don't consider themselves a writer.

The updated agent makes this easier.

Everyone in our Slack workspace can mention @mintlify and have a PR in just a minute. The describe what they shipped, why it's important, and how people use it. Or just link to the PR that shipped the feature. If there are any relevant screenshots or files, they can attach them and the agent will process them. Then they tag me to review the PR the agent creates.

The result: my team contributes more because the barrier is lower, and the docs are better because we're capturing knowledge from the people who built the features.

I create tickets for small documentation updates all the time. Screenshots go out of date. Code samples have a spacing issue. A description could be more clear. For lower priority problems, these tickets could sit in the backlog for weeks or months. But these little things add up to a less polished documentation experience.

Now, instead of creating a ticket, I just @mintlify in Slack and describe what needs updating. The agent creates a PR, I review it, and it's done.

Small improvements ship faster. The backlog is shrinking and stays focused on strategic work rather than accumulating polish tasks.

The key improvements that enable these workflows are:

File and image processing: The agent can now work with attachments from Slack. Someone can send a screenshot or diagram and ask for it to be added to specific documentation. The agent handles the upload and placement.

Multi-file updates: When a change affects multiple pages, the agent coordinates updates across the repository.

PR review feedback: If someone leaves comments on a PR requesting changes, the agent can address that feedback directly. This creates a tighter feedback loop and reduces back-and-forth.

Better planning: The agent now follows a more explicit workflow: research → plan → write → validate → open PR. This leads to fewer mistakes and less clean up of PRs before they can be merged.

The agent still only makes changes through pull requests. It never commits directly to main. This maintains our review process while reducing the manual work of creating and formatting documentation.

If you're managing documentation for a product, especially with a small team, try the updated agent.