Scenario Blender Export logo

Scenario Blender Export

OrganizationPopular
scenario-labs
scenario-blender-export

Use when exporting a model, character or animation from Blender to a game engine, the web or a 3D printer: FBX, glTF/GLB, USD/USDZ or STL for Unity, Unreal Engine, Godot, Roblox, three.js or Apple AR Quick Look. Also when an export arrives at the wrong scale (100x, 0.01), rotated 90 degrees, facing backwards, with extra leaf bones, without its animation, blend shapes or textures; when an FBX loses its metallic or roughness map; when a generated model needs its real size and its pivot on the floor; when making LODs or collision meshes (UCX_, -colonly) for an engine; when converting a folder of models in batch; or when checking an exported file by re-importing it.

Overview

Publisherscenario-labs
Repositoryskills
Skill namescenario-blender-export
Stars
1.2K
Forks
144
Bundled files
2
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 2 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by scenario-labs on GitHub. Read the source before you install it.

Installation

Install the Scenario Blender Export AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/scenario-labs/skills/tree/main/skills/dcc/blender/scenario-blender-export
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/scenario-labs/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/skills/dcc/blender/scenario-blender-export .claude/skills/scenario-blender-export
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Scenario Blender Export in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Scenario Blender Export on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Scenario Blender Export is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Blender export (engine-ready files)

An export is done when the file has been re-imported and measured, not when the exporter returned FINISHED. Gate the asset first, export with a preset, then round-trip the file. This skill builds on scenario-blender-expert (execution channels, the delivery gate) and imports its scripts, so install the two side by side. If a sibling skill named here is missing from your available skills, load it with skill_get (Scenario MCP) or ask the user to install it (https://github.com/scenario-labs/skills#install); unattended, proceed from tool schemas and flag the gap.

Toolkit: scripts/bx_export.py. Per-engine facts with their sources: references/engines.md.

Quick reference

TargetPresetFileEngine side
Unity, static meshunityFBXimport defaults
Unity, characterunity_skeletalFBXleave Bake Axis Conversion off
Unreal EngineunrealFBXConvert Scene Unit (to centimeters)
Unreal Engine, characterunreal_skeletalFBXone root bone
Godot 4godotGLBimport defaults; Generate LODs in the dock
RobloxrobloxFBX20,000 triangles per mesh
three.js and the webwebGLB, meshopt, WebPMeshoptDecoder; not for Godot
any glTF readergltfGLB
USD pipelinesusdUSDC
Apple AR Quick LookusdzUSDZ, 2048 textures
3D printingprintSTL in millimetersslicer reads millimeters

The Unity and Godot presets were imported into Unity 6000.3 and Godot 4.7.2 and measured; unreal, unreal_skeletal and roblox follow Epic's and Roblox's documentation and pass the Blender round trip only. The measurements live in references/engines.md.

Procedure

  1. Gate. Run bx_audit.deliverable(parts, profile) from scenario-blender-expert and fix every error. Scale left on an object travels into the engine as the node's transform.
    • Open borders and edges shared by three or more faces have no automatic fix. On a generated chest, hole filling closed 4 of 24 borders and added a back-to-back face.
    • Rebuild the surface with scenario-blender-retopology, bake the original's maps onto it, then gate again.
    • Without a rebuild, report the file as blocked for that profile (roblox requires watertight meshes). Never deliver it as ready.
  2. Size and pivot. A generated mesh arrives at an arbitrary size with its origin at its center, so it sinks into the floor. bxe.ground(parts, width=0.9) (or height=, in meters) bakes the transforms, puts the center of the base at the origin and scales to the real size. Ask for the size when the brief does not give it; unattended, pick a plausible one and state it as an assumption. Static meshes only.
  3. Pick the preset from the table. Pass every object the file needs: a skinned mesh AND its armature, LOD meshes and their group, collision meshes beside their render mesh.
  4. Export: bxe.export(path, preset, objects, **overrides). Any exporter keyword overrides the preset, such as bake_anim=True on roblox for an animated file. The result records the settings used.
  5. Round-trip: bxe.round_trip(path, objects, preset=...) re-imports the file in a background Blender, leaving the open scene alone. It compares size (1 mm), triangles, deforming bones, root bones (one expected), armature scale, shape keys, every action stored for the objects, and the image maps on each material (base color, metallic, roughness, normal, emission). Then it lists every difference. Fix each one, or say why it is expected.
  6. glTF contents: bxe.glb_summary(path) lists extensions_required, morph targets, skins, clips and image types without Blender. A loader refuses a file that requires an extension it lacks.
  7. Engine check: import into the engine and look. For Godot, the headless check prints each node's rotation and scale, each mesh's size and floor height, and the maps each material received. Full flows: scenario-unity-pipeline-automation, scenario-unreal-pipeline-automation, scenario-godot-pipeline-automation.

Scale, axes and units

  • Blender works in meters, Z-up, with the asset's front facing -Y. glTF is meters, Y-up (export_yup). FBX stores its axes as metadata; the presets write -Z forward, Y up.
  • Keep scene.unit_settings.scale_length at 1.0. Any other value rescales every export.
  • Apply object scale before export: obj.data.transform(Matrix.Diagonal(obj.scale).to_4x4()), then obj.scale = (1, 1, 1), so the engine node imports at scale 1. bxe.ground() does it for every mesh at once.
  • Engines read 1 unit as 1 meter and place an object by its origin. A prop's origin belongs at the center of its base, so it stands on the floor where it is placed (step 2).
  • FBX scale: unity writes FBX_SCALE_ALL (Unity reads file scale 1). unreal keeps FBX_SCALE_NONE (Unity reads it as file scale 0.01, a centimeter file). roblox uses FBX_SCALE_UNITS, as Roblox's export requirements ask.
  • STL has no unit and slicers read millimeters, so print multiplies by 1000.

Characters

  • Export the armature with its meshes. use_armature_deform_only leaves control bones out, so keep every deforming bone in one hierarchy under one root bone. Unreal's importer reports "Found multiple roots. The importer only supports one root per mesh."
  • add_leaf_bones is off in every preset. Left on, each chain gains an extra end bone that breaks retargeting.
  • unity_skeletal and unreal_skeletal bake every action (bake_anim_use_all_actions), one clip per action. Unity names each clip Armature|Action (measured: Rig|Wave), so name actions as the clips should read. glTF exports each action as an animation under its own name.
  • roblox exports no animation. For an animated file, pass bake_anim=True to export() and to round_trip(), which then compares the actions too.
  • Shape keys travel in FBX (blend shapes), glTF (morph targets, names in extras.targetNames) and USD. Building ARKit or viseme sets is scenario-blender-rigging's work.
  • Roblox: at most 4 influences per vertex, bones frozen (scale 1, rotation 0), root at the origin with no influences, one animation track per mesh. Check influences with max(len(v.groups) for v in obj.data.vertices).

LODs and collision

  • bxe.lods(obj, ratios=(0.5, 0.25)) renames the mesh Name_LOD0 and adds decimated Name_LOD1..n under an empty called Name, which takes the mesh's place under its parent. Static meshes only: it refuses a skinned mesh. Unity builds the LOD Group from those names. Godot generates LODs at import (Generate LODs). Unreal documents "Import LODs", all LODs sharing one pivot. The group exports as a plain FBX Null, and Unreal's import of it is unmeasured. It welds the vertices a glTF import splits at UV seams before decimating; without the weld, the LODs open into cracks.
  • bxe.collision(obj, engine, shape, max_verts=64) adds a collision mesh named for the importer:
    • Unreal: UCX_Name_00 (convex hull) or UBX_Name_00 (box), exported in the same FBX. Unreal skips Auto Generate Collision when the file carries one.
    • Godot: Name-convcolonly for convex (one hull) and box (the bounds, a box collider), or Name-colonly for mesh (the mesh as a trimesh, static bodies only).
    • Unity and Roblox: returns None. Colliders are set in the editor.
    • A convex hull keeps at most max_verts vertices, spread over the shape. A generated mesh's full hull runs to thousands of faces. 64 is this toolkit's budget, not an engine limit. None keeps every vertex, and a dense full hull can come out with open edges.
  • Decimation can wreck UV seams and silhouettes. Review every LOD with bx_review before export.

Batch conversion

bash
blender -b --factory-startup --python-exit-code 1 -P bx_export.py -- \
        --batch /abs/in /abs/out --preset godot --profile game

For every model in the folder (GLB, glTF, FBX, OBJ, USD, STL, PLY): import into an empty scene, run the delivery gate, export with the preset, re-import and compare. It writes /abs/out/report.json with each file's gate errors, size and round-trip differences. One failing file never stops the folder: it gets a failed entry. Two inputs sharing a name keep their format in the output name (hero_fbx.glb, hero_glb.glb). Open the report before calling the batch done.

Worked example: a generated GLB to Unity and Godot

python
import sys
sys.path.append("/abs/skills/scenario-blender-expert/scripts")
sys.path.append("/abs/skills/scenario-blender-export/scripts")
import bpy, bx_audit, bx_export as bxe

bpy.ops.wm.read_factory_settings(use_empty=True)
bpy.ops.import_scene.gltf(filepath="/abs/in/robot.glb")
parts = [o.name for o in bpy.context.scene.objects if o.type == "MESH"]
gate = bx_audit.deliverable(parts, profile="game")
print(gate["errors"])            # fix each one, then gate again until "pass" is True
bxe.ground(parts, height=1.2)    # the robot's real height in meters, base center at the origin

bxe.export("/abs/out/robot.fbx", "unity", parts)
print(bxe.round_trip("/abs/out/robot.fbx", parts, preset="unity")["diffs"])
# ["Robot: material Body lost its Metallic, Roughness maps"]: write them for the engine
print(bxe.extra_maps(parts, "/abs/out", "unity")["written"])
bxe.export("/abs/out/robot.glb", "godot", parts)
print(bxe.glb_summary("/abs/out/robot.glb")["extensions_required"])            # []

A generated GLB often holds one object per part, all sharing the asset's pivot. Keep them as separate objects when the engine should animate or swap parts. Join them (bpy.ops.object.join with the parts selected) for a single static prop, then gate again.

Common mistakes

  • Calling an export done on FINISHED. The exporter succeeds on files the engine imports at 100x scale or without the skin.
  • Exporting a skinned mesh without its armature, or the armature without its mesh.
  • Setting Bake Axis Conversion or Apply Transform to remove the armature node's rotation: neither removes it, and the first turns the character around.
  • The web preset for Godot, or any compressed GLB for a loader without its decoder. Check extensions_required first.
  • FBX textures left as external paths: the presets copy and embed them (path_mode="COPY", embed_textures=True), so the file travels alone. export() gives packed images without an extension (as a GLB import leaves them) a .jpg or .png name first. An image never saved (a new bake target) or edited since travels as a packed copy of its pixels (packed_unsaved_images).
  • A textured FBX that shows untextured in Unity: the default import keeps embedded textures inside the file. Extract Textures (Materials tab) into the model's folder assigns them.
  • Trusting an FBX made from a glTF import for metal and roughness. The glTF importer routes one packed image through Separate Color, and the FBX exporter drops both maps: the file re-imports with Metallic 0 and Roughness 0.5. Base color, normal and emission travel. round_trip() reports the loss. bxe.extra_maps(parts, out_dir, engine) writes the two maps as files to assign:
    • Unity: <material>_MetallicSmoothness.png in the Metallic Map slot (Standard or URP Lit), Smoothness Source on Metallic Alpha, sRGB unchecked. Unity's FBX import assigned only the base color, even from a file carrying both maps (measured).
    • Roblox and Unreal: <material>_Metallic.png and <material>_Roughness.png, 8-bit grayscale. On Roblox, set them as MetalnessMap and RoughnessMap of a SurfaceAppearance. In Unreal, uncheck sRGB on both.
    • Godot: deliver the godot GLB, which keeps every map.
  • An STL at Blender scale: a 10 cm part prints at 0.1 mm.
  • LODs exported without a look: decimation can collapse thin parts.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Scenario Blender Export AI skill do?

Use when exporting a model, character or animation from Blender to a game engine, the web or a 3D printer: FBX, glTF/GLB, USD/USDZ or STL for Unity, Unreal Engine, Godot, Roblox, three.js or Apple AR Quick Look. Also when an export arrives at the wrong scale (100x, 0.01), rotated 90 degrees, facing backwards, with extra leaf bones, without its animation, blend shapes or textures; when an FBX loses its metallic or roughness map; when a generated model needs its real size and its pivot on the floor; when making LODs or collision meshes (UCX_, -colonly) for an engine; when converting a folder...

Why use Scenario Blender Export on TypingMind?

Because you install it once and use it with any model. Scenario Blender Export is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Scenario Blender Export in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/scenario-labs/skills/tree/main/skills/dcc/blender/scenario-blender-export. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Scenario Blender Export?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Scenario Blender Export?

As many as you like. As long as a model supports skills, you can use Scenario Blender Export with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Scenario Blender Export AI skill free?

Yes. It is published on GitHub by scenario-labs under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇