fabprint 0.1.71

Headless 3D print pipeline: arrange, orient, slice, and send to printer from a TOML config

fabprint

3D prints are hard to reproduce:

• Slicer settings get lost between sessions
• Printer configs drift across machines
• There's no easy way to version or diff a print job

fabprint makes 3D printing reproducible. Define your models, slicer settings, and printer config once in a TOML file, then arrange, slice, and print from the command line — identically on any machine. It works with STL, STEP, and 3MF files, and pairs naturally with code-CAD tools like build123d, OpenSCAD, and cadquery.

# fabprint.toml — a multi-part print with slicer overrides

[[parts]]
file = "enclosure_base.step"
orient = "flat"
filament = 1                    # AMS slot 1: PETG-CF

[[parts]]
file = "enclosure_lid.step"
orient = "upright"
filament = 1

[[parts]]
file = "button_cap.stl"
copies = 4
filament = 2                    # AMS slot 2: PLA

[slicer]
engine = "orca"
version = "2.3.1"               # pinned for reproducibility
printer = "Bambu Lab P1S 0.4 nozzle"
process = "0.20mm Standard @BBL X1C"
filaments = ["Generic PETG-CF @base", "Generic PLA @base"]

[slicer.overrides]
sparse_infill_density = "25%"
enable_support = 1
brim_type = "auto_brim"

[printer]
name = "workshop"

fabprint run        # arrange → slice → print, one command

How it works

• Everything is text — one TOML config per project, git-friendly and diffable
• Pinned profiles — lock exact slicer, filament, and process profiles in your repo
• Slicer overrides — tweak support, bed type, wall count without touching profile files
• Versioned Docker slicing — pin OrcaSlicer version for identical G-code across machines
• One command — fabprint run goes from STL/STEP files to a running print

How is this different from OrcaSlicer CLI?

This builds on OrcaSlicer CLI, but is designed to allow other slicers like Cura to plugin.

OrcaSlicer CLI slices one plate of pre-arranged models. fabprint is a pipeline around it:

• Arrangement — bin-packs multiple STLs onto the build plate (OrcaSlicer CLI has no arrange step)
• Multi-part filament mapping — per-part filament slot assignment and paint color preservation, injected into the 3MF metadata
• Reproducible builds — pin slicer profiles into your repo + lock OrcaSlicer version in Docker = identical gcode on any machine
• Partial execution — --until plate to inspect layout, --only slice to re-slice, --dry-run to test everything
• Send to printer — Bambu LAN, Bambu Cloud, and Moonraker/Klipper (experimental), with live status monitoring. PrusaLink and OctoPrint support is on the roadmap
• Headless Docker slicing — no GUI, no display server, works in CI

Quick start

Prerequisites: Python 3.11+ and either Docker or a local OrcaSlicer install. Docker is required for Bambu Cloud printing and recommended for reproducible slicing — it lets you pin the exact OrcaSlicer version so every machine produces identical G-code.

pip install fabprint                # STL + 3MF support, LAN + cloud printing
pip install "fabprint[step]"        # add STEP file support (build123d)

Generate a config with the interactive wizard, or dump a commented template:

fabprint init                       # interactive wizard — discovers profiles and CAD files
fabprint init --template            # dump a commented template to stdout
fabprint init --template > fabprint.toml   # save template and edit manually

Or create fabprint.toml by hand (see full config reference):

[pipeline]
stages = ["load", "arrange", "plate", "slice", "print"]

[printer]
name = "workshop"       # references ~/.config/fabprint/credentials.toml

[plate]
size = [256, 256]       # build plate dimensions in mm
padding = 5.0

[slicer]
engine = "orca"
version = "2.3.1"       # pin OrcaSlicer version for reproducibility
printer = "Bambu Lab P1S 0.4 nozzle"
process = "0.20mm Standard @BBL X1C"

[slicer.overrides]
sparse_infill_density = "30%"       # stronger infill
wall_loops = 3                       # extra wall strength
enable_support = 1
brim_type = "auto_brim"             # help adhesion
curr_bed_type = "Textured PEI Plate"

[[parts]]
file = "frame.step"
rotate = [180, 0, 0]    # flip so mounting plate faces down
filament = "Generic PETG-CF @base"

[[parts]]
file = "wheel.stl"
copies = 5
orient = "upright"
filament = "Generic PETG-CF @base"

Run it (see full CLI reference):

fabprint run                   # arrange, slice and send to printer
fabprint run --until plate     # stop after plating
fabprint run --until slice     # stop after slicing
fabprint run --dry-run         # full pipeline without sending to printer

The plate stage generates a plate_preview.3mf — open it in any 3MF viewer to check placement:

Reproducibility

Pin profiles into your repo so builds are identical across machines:

fabprint profiles pin          # copies slicer profiles into ./profiles/
git add profiles/              # commit to lock them

Combined with version = "2.3.1" in [slicer] (which pins the Docker image), the same config always produces the same gcode.

CI/CD example

Automate slicing in GitHub Actions — push a commit, get G-code as a build artifact with print metrics on your PR:

# .github/workflows/slice.yml
name: Slice
on: [push, pull_request]
jobs:
  slice:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: pzfreo/fabprint@main
        with:
          orca-version: "2.3.1"

The action slices your model, uploads G-code as an artifact, and posts print time / filament stats as a PR comment. See action/README.md for all options.

CLI overview

fabprint init                        # interactive config wizard
fabprint init --template             # dump commented TOML template
fabprint validate                    # check config for issues
fabprint setup                       # set up a printer (credentials + connection type)
fabprint run                         # full pipeline
fabprint run --until plate           # stop after plating
fabprint run --only slice            # run just one stage
fabprint run --dry-run               # everything except sending to printer
fabprint watch                       # live printer dashboard
fabprint status                      # query printer status
fabprint profiles list               # list available slicer profiles
fabprint profiles pin                # pin profiles for reproducible builds

Credentials

Printer credentials are stored in ~/.config/fabprint/credentials.toml, created by fabprint setup. The file is set to 600 permissions (owner read/write only) and is never committed to your repo — only the printer name appears in fabprint.toml. Credentials can also be supplied via environment variables (BAMBU_PRINTER_IP, BAMBU_ACCESS_CODE, BAMBU_SERIAL) for CI or shared environments.

Documentation

• CLI reference — all commands, flags, and pipeline stages
• Config reference — complete TOML format
• Developing — setup, testing, architecture

License

Apache 2.0