We're introducing ChatBotKit Studio, a private, native macOS workspace for building and running the ChatBotKit AI platform. It is the same platform we open-sourced last week, packaged as a Mac app: download it, open it, add a model provider, and build bots, datasets, skillsets, portals and integrations on your own machine.
Studio embeds its own verified runtime and starts the platform in an app-private VM. There is no Docker Desktop to install, no Compose file to edit and no ports to juggle. The workspace answers only on loopback addresses on your Mac, and your credentials, conversations and data stay in the app's private storage rather than on a server you share with anyone else.
Managed From the App
Everything you would otherwise do from a terminal lives in native macOS surfaces:
- Settings holds your model provider keys, so the platform starts with the providers you already use.
- Logs and stack details show what the runtime is doing, and a built-in Web Inspector helps when you are debugging a widget or a portal page.
- Multiple windows share one runtime, so you can keep a bot designer and a conversation view side by side.
- Updates are signed and notarized, and the app manages the platform image on your behalf.
Under the hood Studio runs the studio distribution flavor of the platform: SQLite, Redis, Qdrant and Garage on their own port family, so it sits happily beside a Community stack you may already run with Compose. If you would rather drive that flavor yourself, it is published to ghcr.io/chatbotkit/platform-studio and documented in the platform's deployment guide.
Requirements
Studio needs an Apple silicon Mac running macOS 26 or later. The first launch downloads the platform image, so it needs an internet connection and some free disk space; after that the platform runs locally.
ChatBotKit Studio is open source under the Apache License 2.0 and lives at github.com/chatbotkit/studio. Grab the latest release, move Studio.app to Applications, add a model provider in Settings, and start building.