Curious: Thoughts on Adaptive Cards vs. json-render?

[XavierGeerinck]

Hey everyone!

First off, I’m a huge fan of this repo. The work being done here is amazing, and I really love the fluency and "cleanness" of the json-render approach—it feels much more intuitive than other things I’ve used.

I was curious if the team has looked into [Microsoft’s Adaptive Cards](https://adaptivecards.microsoft.com/designer) standard at all? Since they’re a pretty big player in the "UI via JSON" space, I’d love to hear your thoughts on how json-render compares.

I personally think this project is much nicer to work with, but since standards are always helpful for adoption, I wondered if there might be any interesting room for collaboration or alignment there?

Either way, keep up the great work! Excited to see where this goes.

[ctate]

Thanks! Really appreciate the kind words.

Yes, we're aware of Adaptive Cards — it's a mature, well-adopted standard (Teams, Outlook, Windows, etc.). It's great for card-style UIs like notifications, forms, and task panes.

How json-render differs:

Aspect	Adaptive Cards	json-render
Scope	Cards/widgets	Full application UIs
Philosophy	Standard format everyone implements	Flexible format you define
Streaming	Not built-in	First-class JSONL patches
Actions	Submit-style	Rich client-side handling

Our take on standards:

Standards are valuable for interoperability — and we'll support them where it makes sense. But for AI-generated UI, we've found that different teams want different input formats. Some want verbose, explicit JSON. Some want minimal, clean structures.

json-render prioritizes giving developers the freedom to define their own input contracts. You're not locked into one standard — you choose what works for your app.

If Adaptive Cards support makes sense for a use case (e.g., outputting to Teams), we're open to it as an output adapter. But our core philosophy is flexibility over conformity.