usage

Using awan from your code

awan is a binary plus a text protocol, not a library you link. Anything that can spawn a process and write a line can embed it — no SDK.

These are the embed examples. The character most people meet is the one on a profile README; this is the same engine, called from your own tool.

Each folder is a self-contained, runnable example for one language, showing the same use case: keep a companion on screen while a slow task runs, then react to the result.

Language	Run it	You write
node	cd node && npm install && npm start	awan.busy(…), awan.react(…)
python	cd python && pip install -r requirements.txt && python main.py	awan.busy(…), awan.react(…)
go	cd go && go run .	awan.Busy(…), awan.React(…)
bash	cd bash && ./run.sh	awan busy …, awan react …
rust	cd rust && cargo run	Stage::show(…).frame(…)

"Wait — is it just running a binary?"

Under the hood, yes: awan is one small binary. But you never spawn it yourself. The Node, Python and Go packages give you a normal API — awan.react("task.done") — and hide the process boundary completely. In the shell, the awan command is the API, which is as simple as it gets. And in Rust you skip the binary entirely and link the engine (awan-core) directly.

So from your code it reads like any other library — no exec, no spawn, no plumbing.

The events

busy/react take plain event names — cmd.start, cmd.ok, cmd.failed, task.done, idle — and each character's [reactions] decides what they do. Full protocol: docs/INTEGRATE.md.