Pixract: AI Creative Tool with Node-Based Workflows
Pixract is a free, privacy-first AI creative tool that enables users to generate images, 3D models, and text using an intuitive node-based workflow editor.
What is Pixract?
Pixract combines the power of modern AI models with a visual node-graph interface, allowing users to create complex generative workflows. Whether you’re designing images, generating text content, or creating 3D models, the node-based approach lets you connect different AI capabilities and iterate on your creative vision.
All data is stored locally on your device — nothing is stored on external servers. Your creative work stays private.
Key Features
Image Generation
- Create images using Google Gemini (Gemini 3 Pro and Gemini 2.5 Flash)
- Multiple aspect ratios: 1:1, 2:3, 3:2, 3:4, 4:3, 9:16, 16:9, 21:9
- Quality settings: 1K, 2K, 4K
- Reference-based image variations
Text Generation
- Generate text using multiple Gemini models
- Gemini 3 Pro Preview, Gemini Flash Latest, Gemini 2.5 Pro
3D Model Creation
- Generate 3D models using Meshy 4
- Text-to-3D and Image-to-3D conversion
- GLB format support with live 3D preview
Node-Based Workflow Editor
The core of Pixract is its visual node editor with various node types:
- Image Creator — Main image generation node
- Created Image — Stores and displays generated images
- Update Image — Modifies existing images
- Image Reference — Uses images as input for other nodes
- Text Creator — Generates text content
- Text Reference — Uses text as input for other nodes
- 3D Creator — Generates 3D models
- Color Reference — Uses colors as input parameters
Cascade Regeneration
When you modify an upstream node, all connected downstream nodes automatically regenerate. This enables rapid iteration on complex workflows without manually updating each node.
Data Management
- Local Storage: All data stored in IndexedDB (browser-based)
- Image Library: Browse all generated images with full-text search
- Scene Management: Create and manage multiple workspace projects
- Import/Export: Share and backup your scenes
Tech Stack
Frontend:
- SvelteKit 5: For a reactive, modern user interface
- XYFlow/Svelte: For the node-graph visualization
- Three.js with Threlte: For 3D model preview
- Tailwind CSS 4: For utility-first styling
- Bits UI & Shadcn: For pre-built components
Storage:
- IndexedDB: Fully local, browser-based storage
- Repositories for images, nodes, edges, 3D models, and scenes
AI Integration:
- Google Generative AI (Gemini models)
- Meshy API (3D generation)
Hosting:
- Self-hosted on Kubernetes
- Docker-ready deployment
Privacy First
Pixract takes a privacy-first approach:
- No server-side storage of your content
- API calls go directly to AI providers
- Your creative work stays on your device
- You control your API keys
Pricing
Pixract itself is free. Users pay only for the AI generation costs directly to the providers:
| Generation Type | Cost Range |
|---|---|
| Image | $0.03 - $0.24 per image |
| 3D Model | $0.15 - $0.30 per model |
Who is it for?
- Creative Professionals — Designers, artists, and 3D creators
- Content Creators — Bloggers and social media creators needing AI-generated assets
- Privacy-Conscious Users — Those concerned about data privacy with cloud-based AI tools
- Budget-Conscious Creators — Free tool, pay only for API usage