Shipping new features used to mean building the functionality, then stopping to draft docs from scratch, cross-check existing pages, create a pull request, and repeat the cycle whenever inconsistencies surfaced.

We recently shipped the mintlify agent in Slack. It understands your existing documentation, so you can simply tell it what to update and it will generate the pull request for you. You can add extra context to refine its output and keep iterating until it’s exactly right.

Now, the agent will surface suggested documentation changes automatically based on new code you ship, right from the mintlify dashboard.

With the agent you can write new content from prompts or PRs, revise outdated code examples and API references, search and update existing docs, and answer questions about your documentation or technical writing.

How it works

In the dashboard

Open the agent panel from your dashboard with ⌘+I (Mac), Ctrl+I (Windows/Linux), or by clicking Ask agent. Inside, you’ll find:

• Chat for sending prompts and getting proposed doc changes
• History to revisit past conversations
• Settings to manage GitHub and Slack integrations

When the agent proposes updates, you can open the pull request directly from the chat or preview the changes in the web editor.

In your Slack workspace

You can also use the agent directly in Slack to collaborate on documentation updates. If your workspace requires admin approval for apps, make sure the mintlify app is approved first.

To connect:

1. Open the agent panel → Settings
2. Click Connect under Slack integration
3. Add the mintlify app and link your account
4. Test it by sending the agent a DM or mentioning @mintlify in a channel

What's next for automation in documentation

No matter how hard teams try, human-maintained documentation inevitably drifts out of date. The people with the most context have the least time, and those with time often lack the context.

We’re now entering an era of self-updating knowledge systems: documentation that evolves on its own by monitoring pull requests, capturing content from Slack and support tickets, synthesizing updates from multiple sources, and proactively flagging contradictions or outdated information.

This means your knowledge becomes a living source of truth, always up-to-date for both human readers and the AI systems that increasingly rely on it.

Stay tuned for more updates.