The world's first open-source multimodal creative assistant

This is a substitute for Canva and Manus that prioritizes privacy and is usable locally.

中文版| 新手指南

Magic Canva!

"Build" your ideas like playing with LEGO—paint directly, point with arrows, and the AI instantly understands and generates results.

Magic video!

Create Viral Shorts with a Single Sentence

We're evolving Jaaz.app from a traditional "AI tool" into a true "creative partner," redefining human-AI creative collaboration through five core innovations:

• Creative Intention Prediction: AI proactively suggests next steps based on your behavior patterns
• Style Guardian: Automatically detects and alerts style deviations, ensuring consistency
• Creative Catalyst: Proactively provides variation suggestions and cross-media inspiration
• Multimodal Connections: Smart suggestions for pairing images with music, videos with copy

• Multi-dimensional Communication: From task execution to creative collaboration, tactical advice to strategic planning
• Context Awareness: Recognizes your creative state and emotions, adapting response style
• Creative Journey Companionship: Remembers the entire creative process, providing continuity
• Critical Thinking Partner: Proactively questions design decisions, explores better alternatives

• Local Learning Engine: Learns your creative preferences entirely on your device
• Personal Creative Genome: Forms unique style profiles with zero data transmission
• Skill Growth Tracking: Identifies creative weaknesses, provides personalized learning paths
• Cultural Adaptation: Intelligently adjusts suggestions based on your cultural background

• Project Ecosystem Memory: Maintains creative context and project associations across sessions
• Style Evolution Tracking: Visualizes your creative style growth trajectory
• Inspiration Archive: Locally stores creative fragments, reactivating them when relevant
• Creative History Recall: "Should we continue with the poster style we discussed last time?"

• Creative Decision Trees: Clearly displays the logical chain of each design decision
• Style Gene Maps: Visualizes the basis and impact factors of style choices
• Hypothetical Exploration: Interactive "What if we changed to warm colors?" scenarios
• Confidence Indicators: Clearly marks the certainty level of AI suggestions

Star us, and you will receive all release notifications from GitHub without any delay!

🎬 One-Prompt Image & Video Generation Turn one prompt into complete images or videos in seconds.

-Supports GPT-4o, Midjourney, VEO3, Kling,veo3,seedance etc.

-Auto-optimized prompts & multi-turn refinement

🧙 Magic Canvas&Magic Video Prompt-free creation — build like Lego.

-Simple sketching and free combination — AI instantly understands and generates.

-AI understands and generates instantly

-No prompt writing needed

-Describe steps simply on the video, and AI will generate following them.

🖼️ Infinite Canvas & Visual Storyboarding Plan scenes with an unlimited canvas

-Link layouts, manage media visually

-Real-time collaboration supported

🤖 Smart AI Agent System -Chat to insert objects, transfer styles, control logic

-Works with local (ComfyUI) & cloud models

-Maintains multi-character coherence

⚙️ Flexible Deployment & Local Assets -Fully offline or hybrid setup (Ollama + APIs)

-Built-in library for media & prompts

-Cross-platform: Windows & macOS

🔐 Privacy & Security -Local-first, no data leaves your device

-Open-source, no tracking

-Safe for commercial use — you own your data

Download here: https://jaaz.app/

Click the "Log In" button at the top right of the homepage to access API models. With a low-cost plan, you can seamlessly use a variety of powerful APIs.

Start chatting with agent to generate stories or storyboards!

• Prompt: Help me place this character in six different scenes, all in front of landmark buildings from around the world. The lighting is harmonious. He takes photos from all over the world, realistic, with warm light, high picture quality, and a picture ratio of 9:16

• ✅ Privacy-First Personalization: Establish local data storage architecture
• ✅ Experience Continuity Foundation: Implement basic user preference learning and creative journey recording

• ✅ Interaction Sophistication: Implement context-aware conversation and creative collaboration
• ✅ Agency Enhancement: Develop behavior pattern analysis and proactive suggestion systems

• ✅ Cross-media Suggestion Engine: Perfect style consistency detection and multimodal connections
• ✅ Creative Ecosystem: Establish project associations and style evolution visualization

• ✅ Transparent Explainable AI: Implement creative decision trees and style gene maps
• ✅ Enhanced User Control: Perfect confidence indicators and hypothetical exploration tools

• 🔒 Privacy Protection: All personalized learning occurs entirely locally with zero data transmission
• 🚀 Progressive Enhancement: Gradual improvements on existing stable architecture ensuring user experience continuity
• 🧠 Deep Understanding: Upgrade from simple command execution to creative intent understanding and collaboration
• 🎨 Creative Partnership: Not just a tool, but an intelligent partner that understands your creative style
• 📈 Continuous Growth: AI and user grow together, forming unique creative collaboration relationships

🟠 Need Python version >=3.12

First git clone this repo:

git clone https://github.com/11cafe/localart

cd react

npm install --force

npx vite build

cd ../server

pip install -r requirements.txt

python main.py

🟠 Need Python version >=3.12

VSCode/Cursor Install Extensions：

• Black Formatter by ms-python (ms-python.black-formatter)

cd react

npm install --force && npm run dev

cd server

pip install -r requirements.txt

python main.py