Skip to main content

Getting started

Requirements​

  • A Data Link–capable Graphtec cutter (FC9000, CE7000, or newer grit-rolling model) reachable on your network, TCP port 9100.
  • Docker, or Python 3.11+ for a bare-metal run.

Install and run​

git clone https://github.com/Fieldkit-Print/open-graphtec-server.git
cd open-graphtec-server
cp .env.example .env
# edit .env: set CUTTER_HOST to your cutter's IP, DLS_ENABLED=true
docker compose up --build -d

Open http://localhost:8080/ — the status page shows the server, your cutter(s), recent activity, and loaded jobs. On startup the server reads the cutter's model and GP-GL step size and warns if the configured scale doesn't match the machine.

To keep the data folders on the host instead of a named volume, use docker compose -f docker-compose.local.yml up --build -d (paths under ./data).

Bare metal​

cd server
python3 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
CUTTER_HOST=192.168.1.50 DLS_ENABLED=true uvicorn app.main:app --port 8080

Cutter panel setup (one-time)​

  1. Connect the cutter to the LAN and note its IP address.
  2. In the LINK / Data Link menu, set the destination to SERVER (LAN).
  3. Set ARMS to MARK TYPE 2, 4-point (matches the marks this server prints).
  4. Command setting GP-GL or AUTO (AUTO recommended).

Your first barcode cut​

  1. Produce a print PDF with the cut paths marked — see Print preparation. For a quick test you can use the generators in test-sheet/.
  2. Send it through the server: 
    curl -X POST http://localhost:8080/print/prepare \
         -F "file=@labels.pdf" -o labels_marked.pdf -D headers.txt
    The X-Barcode header tells you the assigned code; the cut job is already registered.
  3. Print labels_marked.pdf at 100% scale — never "fit to page".
  4. Load the sheet: printed side up, barcode edge toward the operator, the solid black Start Mark block on the right.
  5. Start the barcode scan from the cutter panel. The cutter reads the code, pulls the job from the server, scans the four corner marks, and cuts along your contours.

Watch it live on the status page — the cutter tile walks through the Data Link states and the activity feed logs the barcode request and the job handoff.

Where things live​

Path (in APP_DATA_DIR, /data under Docker)Purpose
jobs.dbSQLite job library (jobs are stored complete, GP-GL included)
inbox/print/hot folder: print PDFs to prepare
outbox/print/prepared print PDFs (<name>_<barcode>.pdf)
inbox/cut/hot folder: cut-only PDFs (no print output)
processed/consumed input files, timestamped
errors/rejected files + <name>.error.txt with the reason