Questions about Flowbun
What is Flowbun?
Flowbun is an open-source, flow-based home automation runtime where blocks are plain async TypeScript modules stored as files, and wiring is kept as diffable JSON on disk. Every flow is type-checked by the TypeScript compiler before it runs, so miswired connections fail at compile time rather than at runtime. The runtime integrates with Home Assistant, isolates effects to dedicated boundary blocks, and executes each flow in its own process. State such as debounce timers is persisted in SQLite to survive reloads and restarts.
How does Flowbun handle Home Assistant integrations?
The runtime is designed to integrate with Home Assistant, with effects isolated to dedicated boundary blocks and each flow executing in its own process. A crash in one flow never takes down another or your house. Only @hass/* blocks can touch Home Assistant, ensuring that effects live strictly at the boundary of the system.
How does the type-checking system prevent errors?
Blocks are just TypeScript Named, typed inputs and outputs on a plain async function. If you need a switch statement, you write a switch statement — there is no expression language or palette of generic nodes. Wires fail at compile time because every flow compiles through a generated assertion file with real tsc, and a wire carrying the wrong shape fails the reload.
How are flows stored and managed on disk?
Wiring lives beside blocks as plain JSON — diffable, greppable, and editable in your IDE. The runtime only ever reads it, and every write from any tool is auto-committed to its own git history. A block is a file that lives in data/blocks/, one TypeScript module each, while the editor is optional sugar.
How does Flowbun handle state persistence?
State such as debounce timers is persisted in SQLite so it survives reloads and restarts. The example code shows a debounce block that checks the last timestamp against the current signal to prevent rapid state updates, ensuring stability during automation runs.
What tools are available for editing flows?
The editor offers a browser canvas, Monaco with live typecheck errors on save, structured logs filtered per node, and an embedded Claude agent that edits flows through the exact same audited path. All of it is a plain websocket client of the same on-disk files, meaning anything the editor can do, a curl script can do too.
Can you see an example of a real flow?
A real flow, running in a real house, is available at data/wiring/blinds_sun_tracker.json. This demonstrates how the system handles practical automation tasks like tracking sunlight for blinds using the flow-based architecture.
What is the development approach for Flowbun?
The system is described as heavily AI-written but human-reviewed, with a strong emphasis on reading the code before trusting it. The listing description notes that the runtime is designed to integrate with Home Assistant and persists state in SQLite.
How can I get started with Flowbun?
Flowbun is an open-source project, so users can get started by reading the docs and starring the GitHub repository. The platform is free to use as it is open-source, with no stated pricing or subscription plans.
What is the official website of Flowbun?
The official website of Flowbun is flowbun.dev.
How is Flowbun rated on Trustburn?
On Trustburn, Flowbun has a rating of 3.2 out of 5 based on 5 reviews.
Post your question publicly — other visitors, and whoever manages this page, can answer. It will be visible on this page.
Want to hear about this company again? Follow it and we’ll email you when a new review appears.
Based on reviews, listing data and the company's website. Verify anything important with the company directly.
No reviews
Questions & Answers
No questions yet — be the first to ask.
Sign in to post
Your text is saved — you'll come right back to post.
Don't have an account? Create one
News & Press
No news yet. This is where the business shares its own updates and press.
About Flowbun
Flowbun is an open-source, flow-based home automation runtime. Its blocks are written as plain async TypeScript modules stored as files, and the wiring between blocks is kept as diffable JSON on disk. Every flow is type-checked by the TypeScript compiler before it runs, so miswired connections fail at compile time rather than at runtime. The runtime is designed to integrate with Home Assistant, with effects isolated to dedicated boundary blocks and each flow executing in its own process. State such as debounce timers is persisted in SQLite so it survives reloads and restarts.
- Website
- flowbun.dev

