Tiptap AI Toolkit - Empower your AI to directly edit documents in real time.

by
AI Toolkit is now in production-ready beta — and we’re giving 100 builders a free lifetime license. Chatbots are easy. Editing inside a rich-text doc in real time is deceptively hard. Skip months of engineering: Tiptap delivers a safe, reliable bridge between your AI and the documents where your team actually does its work. Sign up for a Tiptap account and join our giveaway: 100 users get a free, lifetime AI Toolkit license (100k tool calls/mo) in exchange for feedback.

Add a comment

Replies

Best

One thing I’d be curious about is how much control developers have over brand voice and personalization. Can the toolkit be configured to follow a company’s tone, terminology and writing rules, or is that mostly left to the model and prompting layer? The editing side looks strong. The bigger test for me would be whether it can improve content without turning it into the usual polished AI fluff. Congrats on the launch!

 Yes, all of this can be customized by configuring the system prompt and context.

The AI Toolkit is a set of tools that you can plug into to your AI agent. Thus, can configure your AI agent using any AI model, context, and system prompt that you want.

 Thanks! Developers have full control over voice and style. This is because the AI Toolkit is just a set of tools that you plug into your AI agent. Therefore, when you define your AI agent, you can choose any AI model, system prompt, or context that you want. You can even combine the AI Toolkit's tools with your custom tools. We want to make it as flexible as possible for developers.

Now this is something I can genuinely use!

 Looking forward to see what you build with it!

Congrats on the launch. The part that stands out to me is the diff engine for structured documents. For AI editing, the trust boundary is usually not the model writing text, but proving exactly what changed and letting humans review it quickly. Are you planning source/intent annotations in the diff view, or is the focus mainly on reliable document operations first?

 Check out our Tracked Changes integration:
Every time the AI makes an edit, it appears as Tracked Changes. Plus, the AI can add an annotation to every edit so that it can explain why it made the edit.

Just upvoted. Dev tools usually have the weakest onboarding emails in SaaS, curious if you're treating that as an edge here.

 Thanks for your upvote. Let me know if you have some tips for onboarding emails.

Rich-text AI gets ugly the moment node identity matters. If an edit moves a table or custom node, does the diff preserve stable node IDs, comments, and plugin metadata, or rebuild that subtree? Text can look right while every external reference quietly breaks.

 This is a good question. If you insert a new node, it will get a new ID, but existing nodes in the document will preserve their existing IDs. Moreover, the AI agent is aware of all the document IDs and metadata and it will try to preserve them when editing.

Congrats on the launch! Real-time collaborative rich-text editing is deceptively hard even without AI in the loop, so a safe bridge for AI to edit live documents is a real unlock. For the 100 free lifetime licenses, what's the one thing you most want feedback on: reliability of the edits, or how natural the AI's edits feel next to a human's?

 We'd like to focus on reliability and developer experience first, so that developers have a good time integrating our tools into their agents. Developers can improve the quality of the responses by adjusting the AI model, model parameters, system prompt and context.