Blender Simulation logo

Blender Simulation

Community
EL4CTEO
blender-simulation

Physics in Blender 5.2 - rigid bodies (density mass, shapes, bake to keys), classic cloth presets and pinning, new node-based Cloth/Hair Dynamics (XPBD, Collider, effectors), Mantaflow fluids, hair curves, GN simulation zones, headless baking. Use for destruction, drapes, flags, falling objects, liquids, smoke, hair motion.

Overview

PublisherEL4CTEO
Repositoryblender-skills
Skill nameblender-simulation
Stars
426
Forks
39
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 EL4CTEO on GitHub. Read the source before you install it.

Installation

Install the Blender Simulation 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/EL4CTEO/blender-skills/tree/main/skills/blender-simulation
  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/EL4CTEO/blender-skills.git /tmp/blender-skills
mkdir -p .claude/skills
cp -r /tmp/blender-skills/skills/blender-simulation .claude/skills/blender-simulation
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Blender Simulation 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 Blender Simulation 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 Blender Simulation 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.

Simulation

The execution model is in blender-core. Procedural sim zones are in blender-geometry-nodes.

One call

scripts/sim_setup.py -- --rigid --active Crate1 Crate2 --passive Floor --density 500 --frames 1 120 --bake --to-keys
scripts/sim_setup.py -- --cloth Flag --preset cotton --pin Pin --colliders Pole --bake
scripts/sim_setup.py -- --gn-cloth Sheet --colliders Table --set Bendiness=0.5 Stretchiness=0.1 --bake

--bake steps through every frame in order (this works headless for point caches and GN simulation zones) and reports final positions and min_z, so you can verify contact. --to-keys turns rigid-body motion into keyframes for engines.

Choose the solver

NeedUse
Falling/stacking/destructionRigid Body (Bullet). Pre-fracture with Cell Fracture (extension) or GN Voronoi
Clothing, flags, drapes (production-proven)Classic Cloth modifier + Collision modifiers
New node-based cloth/hair, custom forces, procedural setupsCloth Dynamics (Experimental) / Hair Dynamics assets (5.2, XPBD)
Liquids, smoke, fireMantaflow fluid domain + flow/effector (fluid.bake_all)
Hair groomsCurves objects + hair node assets (procedural_hair_node_assets), Hair Dynamics for motion
Growth, trails, custom per-frame logicGN Simulation zone
Soft jiggle for gamesbones with spring/jiggle in-engine, or Wiggle-style constraints baked to keys

Rigid bodies

  • Real scale and real mass matter: mass = volume × density (wood 500–700, concrete 2400, steel 7850, plastic 950 kg/m³). The script computes it.
  • Collision shape: CONVEX_HULL for dynamic objects (fast, stable), MESH only for passive concave colliders, and primitives (BOX, SPHERE, CAPSULE) when they fit.
  • Stability: substeps 10 and iterations 10 (set by the script). Use a small collision margin (1 mm). Avoid objects starting interpenetrated.
  • Export: bake to keyframes (--to-keys; the built-in rigidbody.bake_to_keyframes needs a UI, so the script samples matrix_world itself), then export the actions (blender-export-pipelines).

Classic cloth

  • Presets (from Blender's): cotton, denim, leather, rubber, silk. The mesh needs even quad density (2–5 cm edges for clothing). Use a subdivided plane or retopologized garment, never ngons.
  • Pinning: vertex group with weight 1 = pinned (settings.vertex_group_mass). Animate pinned verts with hooks/armature.
  • Collisions: colliders get a COLLISION modifier (thickness 1–5 mm), plus self-collision for folds. Raise quality (steps) for fast motion.
  • Apply cloth as a shape (modifier_apply_as_shapekey) or bake to an Alembic/point cache for render. Games: bake to bones, or apply a static drape.

Node-based dynamics (5.2, experimental)

  • Assets live in datafiles/assets/nodes/geometry_nodes_dynamics_assets.blend: Cloth Dynamics (Experimental), Hair Dynamics, Collider, Set Effector, Custom Force, Custom Effector. The script appends them.
  • Cloth inputs include Pin Group (attribute), Stretchiness, Bendiness, Substeps (5), Constraint Steps (15), Mass, Friction, Collision Radius, Linear Damping, Gravity, Tearing (+ mode/threshold), Effectors Collection, Effectors (bundle).
  • Colliders: put a Collider modifier on each collider object and add those objects to the collection fed into Effectors Collection (the script does both).
  • Hair: Add → Empty Hair sets up the surface + Capture Rest Geometry. Hair Dynamics reads the rest/deformed surface through geometry bundles.
  • Limits (5.2): no self-collision yet, no built-in wind/turbulence (build a Custom Force). Interfaces may change, so pin the Blender version in production.
  • The solver itself (GeometryNodeXPBDSolver) takes a World bundle, Delta Time, Substeps, Constraint Iterations. Building directly on it is advanced; start from the assets.

Fluids (Mantaflow) quick setup

Domain object (FLUID modifier fluid_type="DOMAIN", domain_type="LIQUID"|"GAS"), resolution 64 preview / 128–256 final, cache directory set and cache_type="ALL" for headless; flow objects fluid_type="FLOW" (flow_type LIQUID/SMOKE/FIRE, flow_behavior INFLOW/GEOMETRY); obstacles EFFECTOR. Bake with bpy.ops.fluid.bake_all() under a context override with the domain active. This is slow. Test at low resolution first.

Rules

  • Simulations are frame-rate and scale dependent. Set scene.render.fps and the scale before tuning.
  • Always bake before rendering/exporting, and never rely on viewport playback order.
  • GN simulation zones: bake per modifier (mod.bakes, object.geometry_node_bake_single), or step frames sequentially (script) in background.
  • Point caches in background: stepping frames from frame_start fills the cache. ptcache.bake_all(bake=True) needs a valid context.
  • Verify with look(mode="frames") plus the numbers (min_z ≈ the collider top = contact, no tunnelling).

Gotchas

  • Objects tunnelling through colliders: thin colliders, too few substeps, or too-fast motion. Thicken, raise substeps/quality, or use a bigger collision distance.
  • Exploding cloth at frame 1: starting interpenetration or a zero-area face. Run blender-asset-qa, offset the start.
  • Rigid body world cache ranges are separate from the scene range (the script syncs them).
  • Changing anything (frame range, settings) invalidates caches. Re-bake.

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 Blender Simulation AI skill do?

Physics in Blender 5.2 - rigid bodies (density mass, shapes, bake to keys), classic cloth presets and pinning, new node-based Cloth/Hair Dynamics (XPBD, Collider, effectors), Mantaflow fluids, hair curves, GN simulation zones, headless baking. Use for destruction, drapes, flags, falling objects, liquids, smoke, hair motion.

Why use Blender Simulation on TypingMind?

Because you install it once and use it with any model. Blender Simulation 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 Blender Simulation in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/EL4CTEO/blender-skills/tree/main/skills/blender-simulation. 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 Blender Simulation?

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 Blender Simulation?

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

Is the Blender Simulation AI skill free?

Yes. It is published on GitHub by EL4CTEO 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 👇