Circuitry is the ultimate canvas for AI, workflows, design, and code. On one visual canvas you can whiteboard ideas, design screens, build AI-powered automation workflows and multi-agent systems, write and debug code, generate code in 13 languages, and deploy APIs instantly.

Do I need coding skills to use Circuitry?

No, Circuitry is a no-code platform. You can build complex workflows using our visual drag-and-drop interface. However, developers can also export workflows as code in 13 different programming languages.

What programming languages does Circuitry support?

Circuitry can generate code in 13 languages: JavaScript, TypeScript, Python, C, C++, C#, Go, Java, Rust, PHP, Ruby, Kotlin, and Swift.

Can I create APIs with Circuitry?

Yes, every workflow in Circuitry can be instantly deployed as an API endpoint with webhook support. You get a unique URL for each workflow with enterprise-grade security options.

OpenClaw in Circuitry

Run OpenClaw from a Circuitry chat: in the chat view, in its terminal, or in its own web app, all on one conversation.

You don't need a plugin here. OpenClaw gets your Circuit's models and Circuitry's tools automatically.

Desktop appYour own AI

Add it with Set up AI… in the Ask chip, or Add an agent… in the chat's model picker. Circuitry installs it if it isn't on your computer.

If you haven't used OpenClaw before, run openclaw onboard in a terminal first. It asks for a model and starts its gateway.

Three ways to use it

Chat view: pick OpenClaw as the chat's model. The reply appears when the turn finishes, rather than word by word.
Terminal: the terminal button in the chat header opens OpenClaw's own terminal app.
Web app: the globe button in the chat header opens OpenClaw's web app in a Browser tab, with its own side panels, signed in and on this chat's conversation.

All three share one conversation, so you can start in one and carry on in another.

Circuitry's tools

OpenClaw gets Circuitry's tools automatically whenever a chat or its terminal opens, so there is nothing to connect.

Choosing a model

The model chip under the composer lists what OpenClaw can use. On a phone, or anywhere in the chat or its terminal, type /model.

The model you pick applies to that conversation everywhere: the chat, its terminal and its web app.

In the chat view, /reset starts the conversation over. /status, /models, /think and /help only work in its terminal, and the chat tells you so.

A new chat starts on the model you last picked in Circuitry, or OpenClaw's default if you haven't picked one. A model you picked runs as chosen, so OpenClaw's fallback models don't apply to it.

Models on all your computers

There is nothing to install or pair beyond Circuit itself:

Install Circuit on each computer.
Put them in a cluster: Cluster… in the Circuit menu.

OpenClaw then lists the models your Circuit shares under circuit/: circuit/<model> for this computer's, and circuit/<computer>/<model> for another's, such as circuit/pc/qwen3.8:27b. Pick one with /model. New models appear the next time a chat or terminal opens.

The Circuit plugin is only for OpenClaw running on its own, outside Circuitry.

Good to know

OpenClaw works in its own workspace. Circuitry tells it which project folder you have open, but it keeps its own memory and instructions there.
It has no read-only mode, so it can't run in plan mode or as a reason-only step in a workflow. Circuitry says so instead of running it with full tools.
Its web app opens in a Browser tab in the desktop app. On the web it opens in a new browser window on the computer running OpenClaw. On a phone or another computer it doesn't open yet, so use its terminal or the chat view there.