FollicleFX – GPU-Accelerated Hair Grooming Add-on for Blender 5.0+
Hi everyone,
We’re releasing FollicleFX, a new hair grooming add-on for Blender developed by a small team of engineers and artists. The goal of the project is to provide a production-focused grooming workflow for high-density hair, with performance characteristics closer to traditional studio tools, while staying fully inside Blender.
FollicleFX integrates directly into Blender and uses a native C++ backend for performance-critical operations, with optional GPU acceleration.
Overview
FollicleFX is designed around a non-destructive modifier workflow, inspired by tools like XGen and Ornatrix, but built specifically for Blender’s Curves system.
Key focus areas:
- Interactive grooming at higher strand counts
- Predictable, modifier-based workflows
- Expression-driven control for procedural variation
- Native Blender sculpting and rendering integration
Core Features
Native C++ Backend
- Compiled C++ core for strand generation and modifiers
- Optional CUDA acceleration on NVIDIA GPUs
- Multi-threaded CPU fallback (OpenMP) for non-NVIDIA systems
Workflow
- Non-destructive modifier stack with undo/redo
- Profile curves for root-to-tip control
- Texture maps and painted masks for spatial variation
- Expression-driven parameters using a small scripting language (FollicleScript)
Blender Integration
- Uses Blender’s Curves Sculpt Mode and standard brushes
- Hair updates live with mesh deformation and animation
- Works with Cycles and other render engines via native attributes
- Requires Blender 5.0+ (new curves API)
Modifiers (Current Set)
- Clumping (hierarchical clumps, sub-clumps, noise, tip convergence)
- Frizz (multi-octave noise, flyaways)
- Coil (spiral / helix curls)
- Flow Direction (curve- and arrow-based flow control)
- Cut (mask- and expression-driven trimming)
- Loc / Dreadlock (structured buildup)
- Braid (experimental)
- Simulation (early development)
Expressions (FollicleScript)
Expressions can be evaluated per-strand or per-clump and are compiled for performance.
Examples:
rand(index) # Per-strand random value
lerp(0.3, 1.0, t) # Root-to-tip gradient
rand(clump_id) # Per-clump variation
v > 0.5 ? 1.0 : 0.5 # Conditional logic
map(u, v) # Texture sampling
Performance (Context)
Tested on an RTX 3060:
- ~500k strands generated in under 50ms (GPU)
- Clumping and expression evaluation remain interactive at similar counts
- CPU fallback available, but intended mainly for lighter grooms
Performance varies depending on groom complexity and enabled modifiers.
Hair Cards Generator
FollicleFX includes an in-progress hair cards generator aimed at real-time and game pipelines:
- Automatic card clustering
- Texture baking (diffuse, alpha, normal, flow, AO, ID)
- Multiple LOD levels
- Export to OBJ / FBX / glTF
Note: This feature is still under active development and may require iteration depending on the groom.
Availability
FollicleFX is a paid add-on.
We’re currently offering a Founder Edition during early access:
- Perpetual license
- Commercial use allowed
- All 1.x updates included
Founder price: $79
Planned regular price: $149
Documentation & Tutorials
Written documentation is available and covers installation, core concepts, and modifier usage.
Tutorial videos and guided walkthroughs will be released progressively as development continues. For now, the documentation is the primary reference and is being expanded alongside the tool.
Current Limitations
- Windows only (macOS/Linux planned)
- Blender 5.0+ required
- CUDA acceleration limited to NVIDIA GPUs
- Some features (braids, simulation) are still evolving
Links
Website: https://folliclefx.com
Documentation: https://folliclefx.com/docs
Discord: https://discord.com/invite/sDeYJ9SGZy
We’re actively gathering feedback and are happy to answer technical questions about the approach, limitations, or roadmap.
Thanks for taking a look.
— The FollicleFX Team
