Panda2Prusa 🐼

Bambu Studio → PrusaSlicer, nothing lost in translation.

Convert Bambu Studio / OrcaSlicer .3mf project files into PrusaSlicer-compatible .3mf files, preserving:

• Real per-object transforms (position, rotation, scale) — composed correctly from the build items and component references, not a hardcoded matrix.
• Component / assembly structure — Bambu splits geometry across 3D/Objects/object_N.model and references them via the 3MF production extension; these are resolved and flattened into a single PrusaSlicer-style 3D/3dmodel.model.
• Per-triangle painting (paint_color → PrusaSlicer mmu_segmentation) — the shared TriangleSelector bitstream is preserved (and transcoded when extruders are renumbered).
• Multi-material part / extruder assignments from Metadata/model_settings.config → PrusaSlicer merged-mesh volume config (firstid/lastid ranges).
• Filament → extruder mapping — Bambu AMS slots are sparse (a two-color print may use slots 1 and 3); the converter asks how to map them (or auto-compacts to extruders 1..K) so the model prints on a 2-tool/5-tool Prusa. Paint data is re-encoded to match.
• Multiple plates — pick one or convert all.

This is a ground-up rewrite of the approach in raistlinJ/3mf_bambu2prusa, fixing its core limitations (hardcoded transform, ignored component structure, regex-based XML surgery).

Why the rewrite

The original tool stamped every object with a single hardcoded transform copied from one specific model (a ~0.8× scale at a fixed bed position), ignored Bambu's component/assembly structure entirely, and rebuilt models with fragile string substitution. Any file other than the author's sample came out mis-scaled and mis-placed, with multi-object plates collapsing on top of each other. This version reads the real structure with a namespace-aware XML parser and preserves it.

Install

pip install -r requirements.txt

(Requires Python 3.9+. tkinter ships with the standard Python installer for the GUI.)

Usage

CLI:

python -m panda2prusa input.3mf output.3mf          # convert all plates
python -m panda2prusa input.3mf output.3mf --plate 1 # convert only plate 1
python -m panda2prusa --info input.3mf               # inspect a 3mf without converting
python -m panda2prusa input.3mf output.3mf --map 3=2 # Bambu filament 3 -> Prusa extruder 2
python -m panda2prusa input.3mf output.3mf --keep-extruders  # no renumbering, no prompt

Run interactively and the CLI shows the filaments the file uses (with their colors) and asks how to map them; the GUI pops the same question with color swatches.

Viewing tip: PrusaSlicer only renders multi-color assignments under a printer profile that actually has multiple extruders/tools (MMU3, XL 2T/5T…). Under a single-extruder profile everything legitimately shows as one color.

GUI:

python -m panda2prusa.gui

Standalone Windows exe (no Python needed to run it):

pip install pyinstaller
pyinstaller --onefile --windowed --name Panda2Prusa gui_launcher.py

The result lands in dist\Panda2Prusa.exe — double-click to launch the GUI. Prebuilt exes are attached to the GitHub releases.

Running the tests

The suite runs against any Bambu/Orca .3mf files you drop into samples/ (kept out of git), or a directory pointed at by the B2P_SAMPLES environment variable. Painted, multi-plate, and multi-material files give the best coverage; everything skips cleanly if no samples are present.

pip install pytest
pytest

Status

Working end to end — see CHANGELOG.md for what's in the current release and its known limitations. Issues and PRs welcome, ideally with a sample .3mf that shows the problem.

Support

If this saved your print (or your sanity), you can buy me a coffee. ☕