MCP apps are the new website: they're real, interactive apps that run right inside AI assistants like ChatGPT and Claude. The tooling to build them didn't exist, so we built it. Meet Skybridge, the React framework that handles the infrastructure and the dev loop so you can focus on your features: MCP server, view rendering, client compatibility, testing tunnel, and more. Code once, ship everywhere. Already +500K downloads and dozens of apps in the stores!
Love that it handles MCP server + view rendering out of the box. As someone building inside AI assistants, this removes so much friction. Any plans for more client compatibility (e.g. custom UIs)?
@sarveshsea Hey Sarvesh, yes, definitely :)
We're constantly monitoring and adapting Skybridge to the different clients supporting MCP Apps. The ones specifically on our radar right now are Gemini and Mistral.
The idea that MCP apps could become the "new websites" is really interesting. Curious what the biggest challenge has been in making apps work consistently across different AI assistants like ChatGPT and Claude?
@lera_kuntsevich8 Probably staying up to date with the AI assistants' SDKs and APIs. They've been moving fast these last months, and it has been a real challenge to keep up. I have to say, we pulled it off thanks to our awesome community :)
The interesting bit for me is the compatibility layer across assistants; MCP apps will be much easier to adopt if teams don’t have to maintain separate UI paths for Claude, ChatGPT, and internal agents.
@jimmy_lee12 exactly the problem we’re trying to solve with Skybridge : a single interface and API set for all hosts. The React native of AI Apps. Hosts all have weird quirks that get polished by our implementations. Thanks for the comments and hope you find it nice to build with Skybridge ;)
Love the abstraction layer here. Building standard UIs while the big players fight over becoming an OS is a smart play. I haven't tested it yet, but it's next on my list. I'll give it a spin soon and will definitely ship some apps once Gemini finally joins the party and opens up an app store! 🚀
@cerkiner thank you for your comment! don't hesitate to give your feedback on our Discord or direcly on our Github :)
@cerkiner thanks for giving Skybridge a try! Let us know your feedbacks when you do, and share what you built on our discord!
Gemini spark will be shipping a first closed entry App Store this summer with a first set of 30 apps, but they’ll open up submissions at some point later down the road, like ChatGPT did!
@cerkiner Can't wait to hear your feedback on it. And yes Gemini announced an extended support of MCP at Google IO but it's unclear if MCP apps will be released soon. We will keep you posted.
Building apps in the cosmetics space, Skybridge does really help on improving the overall building experience, especially the devtools! Thanks for putting this together. What’s plan for next release?
@antoine_grimal3 glad to hear you building in the comstic space and that Skybridge helps doing so!
We got a lot coming in the next release, can't share all, but sneak peak includes better auth setup for private servers and evals... 🤫
Hey @julienvallini - the localStorage viewstate shim is smart for Claude's lack of native persistence. What happens to it when someone has two claude conversations open at once, does the store get clobbered, or is it scoped per-conversation?
@taylor_time as far as Claude is concerned, it's scoped per conversation, per view even in the same conversation, so that multiple can be displayed at once!
Will give skybridge a try for Vook! Thanks for sharing this. Pretty sure this can be a good acquisition channel for people asking chatgpt to transcribe for them: Any idea how voice can be integrated in an MCP app experience?
@jeremy_rippert thanks, voice in MCP app is still in early stage. The standard enable media content negotiation with hosts, but only Claude so far enables microphone access. But your use case sounds like something worth trying to implement. Let me know how it goes! Would probably add an exemple app too on the repo to showcase audio demo