Writing flows
Applications and methods
Where the callable methods come from, and how they document themselves.
Applications live in the applications folder of your context directory. Each
one is a Node.js module that exports methods, and each method is what a step
can call. They can talk to HTTP APIs, MQTT, PostgreSQL databases, or drive a
browser with Playwright.
Documentation is read straight from the JSDoc blocks of the application's
index.js — there is no docs.json. That documentation is what the UI renders
and what the model is given when it writes a flow for you, so the better the
JSDoc, the better the generated flows.
| Tag | Meaning |
|---|---|
@param {type} name - description |
An input parameter. [name] marks it optional, [name=value] adds a default. |
@returns {status} description |
The response, with an optional JSON example body. |
@memory {write|read} key - description |
Flow memory the method writes or reads. |
@example |
An example step, in YAML, ready to paste into a step block. |
A README.md in the application folder is rendered in the UI as well.
The tool seeds three example applications on first run: calculator (fully
offline), httpbin and jsonplaceholder.