Building U1 profiles

Where the toolkit looks for slicer profiles, why it ships empty, and how to build per-extruder, per-filament profiles from your own successful prints.

← Back to the README

---

Profile sources

The toolkit no longer ships default profiles. A fresh install has an empty picker. Profiles come from one of three sources you populate yourself, scanned in priority order:

Source dir	Populated by	Purpose	Priority
profiles/from-printer/	python3 tools/extract_profiles_from_printer.py	Profiles extracted from your printer's recent G-code history. Physics-validated — every setting produced a successful print.	Highest
profiles/user/	The operator, manually	Hand-tuned overrides + custom variants you want to keep stable across stock refreshes	Middle
profiles/snapmaker-stock/	python3 tools/fetch_snapmaker_profiles.py	Snapmaker's official U1 profiles, pulled fresh from the Snapmaker/OrcaSlicer upstream repo (~217 files: every nozzle size + layer height + Snapmaker-tuned filament)	Lowest (universal baseline)

All three are listed in .gitignore — they're per-user, not redistributed.

First-run setup

# Pull Snapmaker's official U1 baseline (~217 files, one-time):
python3 tools/fetch_snapmaker_profiles.py

# Extract whatever you've actually printed successfully so far:
python3 tools/extract_profiles_from_printer.py

Both are idempotent — re-run anytime to pick up Snapmaker upstream updates or fresh prints from your printer's history. Snapmaker stock gives you the universal U1 baseline; extracted profiles reflect what you've validated on your hardware.

Without either, the workflow fails closed at analysis time with a clear setup_required event pointing you back here. Hermes agents surface that error verbatim.

Why ship empty

Earlier versions shipped 13 personal community profiles in profiles/ as defaults. They were tuned for one bed surface (Textured PEI), one bed temp, specific filament brands (SUNLU PETG, HF White PETG), specific tool assignments. Running them silently on another U1 with different filaments or a different bed surface could ruin prints — and the toolkit had no way to warn the user that the profile underneath didn't match their setup.

v1.5.0 moves those personal templates to examples/profiles/ and points the picker at three honest sources: Snapmaker upstream (universal baseline), your printer's history (physics-validated on your hardware), and your own hand-tuned profiles. The agent's Preset? prompt now annotates each option with source, has_supports (read from the JSON's enable_support field), and supports_status (does picking "Add supports" auto-promote to a _supports sibling, already encode supports, or fail with a no_supports_variant warning).

Supports auto-detection

Profiles are JSON-typed for supports — the picker reads each profile's enable_support field and annotates the option with has_supports: true/false. The agent's Preset? prompt also carries a supports_status that pre-warns the user before the Supports? question:

• "self" → preset already encodes supports; "Add supports" is a no-op for them
• "<variant_name>" → if user picks "Add supports", workflow auto-promotes to this same-source sibling and emits a preset_promoted event
• null → no same-source supports sibling exists (or multiple ambiguous candidates). Workflow emits a warning event with kind:no_supports_variant and slices without supports — agent surfaces it before the user trusts the preview

Why "same-source exactly one"? Snapmaker stock has multiple Support flavors at the same layer height (0.20 Support, 0.20 Support W, 0.20 Bambu Support W) — auto-promote can't pick one; the user has to.

Why build your own (vs. just importing the examples)

Profile-as-data-from-real-prints means every setting is physics-validated — it produced a completed print on actual hardware. But the validation is environment-specific:

• Different bed surface (smooth PEI, garolite, glass) → different first-layer temp / bed temp / Z-offset
• Different filament brand → different optimal nozzle temp (PETG ranges 230–260°C across brands)
• Different tool assignment → e.g. your PLA is in extruder0, mine in extruder2
• Different exhaust/enclosure → affects warping defaults

Importing someone else's profiles is fine as a starting point; running them as gospel on a different setup will give you mediocre prints.

Build per-extruder, per-filament profiles from your own print history

This is the recipe used to bootstrap the included community profiles. It only takes one good print per filament-type-per-extruder slot.

The fastest path — one command:

python3 tools/extract_profiles_from_printer.py

That connects to your U1 (via SNAPMAKER_U1_HOST / .env), pulls the 5 most recent G-codes, runs the extractor against each, and drops process + filament JSONs into profiles/from-printer/ — with multi-tool metadata sliced down to the actual tool each print used (so the filament profile for a T1 PETG print isn't polluted by T0/T2/T3 settings).

Tweaks: --list to see what's on the printer first, --file "<exact gcode>" to pick a specific one, --limit N to grab more, --vendor SUNLU to override the often-generic vendor field, --output-dir <path> to write elsewhere.

The longhand recipe — same outcome, manual steps:

1. Print once with Snapmaker's defaults — get a clean part, no warping/stringing/under-extrusion, on your bed surface and filament. Just enough to call it "good enough to use as a baseline."

2. List successful prints via Moonraker:

   curl http://YOUR_U1:7125/server/files/list?root=gcodes
   

3. Download the G-code and parse the ; key = value metadata block at the top. The key ones:

   ; filament_type, filament_settings_id
   ; print_settings_id
   ; layer_height, first_layer_height
   ; nozzle_temperature, first_layer_temperature
   ; bed_temperature, first_layer_bed_temperature
   ; curr_bed_type
   ; sparse_infill_density, wall_loops
   ; nozzle_diameter
   

4. Build a flattened process JSON (see examples/profiles/community_merged_*.json for shape) and a matching filament JSON (see examples/profiles/community_generic_petg_*.json). Name them with the extruder + filament so you don't confuse yourself: e.g. myprinter_extruder1_sunlu_black_petg.json.

   Or run the included extractor to do steps 3-4 in one go:

   python3 tools/extract_profile_from_gcode.py my_good_print.gcode \
       --process-out  profiles/myprinter_extruder1_petg_process.json \
       --filament-out profiles/myprinter_extruder1_sunlu_black_petg_filament.json \
       --process-name  "My 0.20 PETG Extruder1" \
       --filament-name "My PETG Extruder1" \
       --vendor SUNLU --brand-label "SUNLU Black"
   

   It parses the slicer's ; key = value metadata block, emits a flat process JSON + a list-shaped filament JSON in Snapmaker Orca's expected shape, and lets you override filament_vendor (G-code often says "Generic"). Pass --metadata-only to inspect the raw parsed keys without writing files.

5. Track per-extruder mapping in u1_tool_material_map.json so the toolmap gate enforces correct slot assignment:

   {
     "tools": {
       "extruder":  { "material": "PLA",   "label": "Polymaker PolyLite Black" },
       "extruder1": { "material": "PETG",  "label": "SUNLU Black PETG" },
       "extruder2": { "material": "PETG",  "label": "HF White PETG" },
       "extruder3": { "material": "PLA",   "label": "Polymaker PolyLite Grey" }
     }
   }
   

The toolmap gate (u1_toolmap.py) then prevents you from accidentally slicing a job for PETG and uploading it against the slot loaded with PLA.

Reference: example community profiles in examples/profiles/

The 13 profiles used during development live in examples/profiles/ as a shape reference. They're MIT-licensed and show what a working community-tuned profile looks like for the U1. Do not use them as defaults — they assume Textured PEI + specific filament brands. If you happen to share that setup, copy them into profiles/user/ and they'll appear in the picker.

The naming convention so you can see the pattern:

Pattern	Meaning
community_016_optimal_*	0.16mm layer, optimal preset, process profile
community_020_strength_*	0.20mm layer, strength preset (6 walls, 25% infill)
community_*_supports	+ tree/auto supports enabled
community_*_gyroid	+ gyroid infill pattern
community_*_fuzzy_external	+ fuzzy skin on outer walls
community_generic_petg_*	Filament profile for PETG
community_*_sunlu_black_*	SUNLU brand-specific (240°C first layer)
community_*_hf_white_*	High-flow white PETG variant
community_merged_*	Flattened — works for headless CLI slicing
community_*_override	Inherits from official — GUI only

Diff against the official Snapmaker preset chain is ~93% identical; deltas are tuning choices that came from real prints (lower prime-tower waste, arachne walls, brand-specific PETG temps).

Use them as templates to copy + modify for your own setup. Don't blindly import.

File	Type	Use case
community_merged_016_optimal_u1_textured_pei.json	process	Start here. Flattened 0.16 Optimal, no inheritance — works headless
community_016_optimal_u1_textured_pei.json	process	Standalone 0.16 Optimal
community_016_optimal_u1_textured_pei_override.json	process	Inherits-from-official override
community_016_optimal_*_fuzzy_external*.json	process	Fuzzy/staggered seam variants
community_020_strength_u1_textured_pei.json	process	0.20 Strength preset
community_020_strength_supports_*.json	process	Strength + supports
community_020_strength_gyroid*.json	process	Strength with gyroid infill
community_generic_petg_u1_textured_pei.json	filament	Generic PETG (255°C first layer)
community_generic_petg_sunlu_black_*.json	filament	SUNLU Black PETG (240°C first layer)
community_generic_petg_hf_white_*.json	filament	High-flow White PETG

Diff against official ≈ 93% identical; deltas are documented tuning choices, not regressions.

Importing profiles into OrcaSlicer (GUI)

1. Open OrcaSlicer → top-right gear → "Configuration / Profiles"
2. Drag-and-drop the desired .json file into the profiles panel, OR copy to the system config directory for your slicer:
   • Upstream OrcaSlicer (recommended): ~/.config/OrcaSlicer/system/Snapmaker/process/ (or filament/) on Linux/macOS; %APPDATA%\OrcaSlicer\system\Snapmaker\process\ on Windows
   • Snapmaker fork (if you're using snapmaker-orca instead): ~/.config/SnapmakerOrca/system/Snapmaker/process/ / %APPDATA%\Snapmaker_Orca\system\Snapmaker\process\
3. Restart OrcaSlicer
4. Select the Community profile from the dropdown when slicing

---

← Back to the README