pygrbl-server 0.0.1

FastAPI router and job engine for streaming G-code to GRBL controllers

pygrbl-server

FastAPI router and thread-safe job engine for streaming G-code to GRBL controllers. A library, not an application: mount the router in your own FastAPI app and run it with the ASGI server of your choice. One active job per serial port, live status, and a retained terminal result per port.

Part of the pygrbl family: pygrbl-build (generate G-code) → pygrbl-streamer (stream to GRBL) → pygrbl-server (expose over HTTP).

Install

pip install pygrbl-server

Dependencies: fastapi and pygrbl-streamer. Bring your own ASGI server.

Usage

from contextlib import asynccontextmanager

from fastapi import FastAPI
from pygrbl_server import JobRegistry, create_router

registry = JobRegistry()  # streamer_factory(port, baudrate), defaults to GrblStreamer


@asynccontextmanager
async def lifespan(app: FastAPI):
    yield
    registry.shutdown()


app = FastAPI(lifespan=lifespan)
app.include_router(create_router(registry))

uvicorn yourmodule:app --host 0.0.0.0 --port 8000

Events are logged to the pygrbl_server logger (alarms and failures at WARNING/ERROR). Inject a custom streamer_factory to swap the streamer backend.

API

Endpoint	Description
POST /stream	Start a job: {"port": "...", "files": [...]}. 202 started, 409 port busy, 404 file missing.
POST /stop?port=	Cancel the active job. 404 if none.
POST /pause?port=&paused=	Feed hold (true, default) or resume (false). 404 if none.
GET /status?port=	Port state; omit port for all ports.

Status is idle, working, paused, or the retained done / error of the last finished job (kept until a new job starts on that port):

curl 'localhost:8000/status?port=/dev/ttyUSB0'
# {"port": "/dev/ttyUSB0", "status": "working", "progress": 42, "error": null}

Security

No authentication. Any client that can reach the port can control physical machines capable of causing fires or injury. Run on a trusted LAN only. The files in /stream are paths on the server's filesystem, not uploads. Auth, TLS, and rate limiting belong in the app that mounts the router.

Design

The library streams files and reports state; everything else lives in your layers:

• Machine setup (homing, air assist, preambles) belongs in your G-code.
• Delays between files: M5 + G4 P<seconds> in your G-code.
• Notifications: poll /status or attach a handler to the pygrbl_server logger.
• Runner, CLI, auth, CORS: your app's job — this ships as a router on purpose.