Frequently Asked Questions About Leonardo.ai
Leonardo.ai (leonardo.ai) is a professional AI image generation platform with proprietary models, LoRA training, and character consistency tools. Below are answers to the most commonly asked questions.
Leonardo Phoenix Model
What is Leonardo Phoenix?
Leonardo Phoenix is Leonardo.ai’s proprietary image generation model, developed in-house rather than based on open-source architectures. It’s designed for professional creative work with emphasis on:
- Strong prompt adherence (following complex, detailed descriptions)
- Visual consistency across generations
- High detail rendering
- Professional-quality output suitable for commercial use
How does Phoenix compare to Stable Diffusion and Midjourney?
- vs. Stable Diffusion: Phoenix offers comparable quality to the best SD checkpoints but through a hosted platform that doesn’t require local GPU hardware or configuration
- vs. Midjourney: Phoenix offers more control and better prompt adherence; Midjourney offers higher average aesthetic quality with less effort
- vs. Flux: Phoenix is comparable in quality; Flux offers open-weight flexibility while Phoenix is proprietary
Can I use Phoenix on the Free plan?
Free plan users have limited access to Phoenix. The model is available but with reduced priority and potentially limited feature access. Full Phoenix access requires the Apprentice or Artisan plan.
Fine-Tuning and LoRA Training
Can I train custom LoRA models on Leonardo?
Yes. Leonardo offers in-platform LoRA training on paid plans (Apprentice and Artisan). The process:
- Upload 10-50+ reference images
- Configure training parameters (Leonardo provides recommended defaults)
- Train on Leonardo’s infrastructure (typically 15-30 minutes)
- Use the trained LoRA in subsequent generations
What can I train a LoRA on?
Common training subjects:
- Characters: Maintain consistent face, body, and features
- Art styles: Capture a specific visual aesthetic
- Products: Render specific products accurately
- Concepts: Teach the model specific visual concepts
- Environments: Capture architectural or environmental styles
How many training images do I need?
- Minimum viable: 10-15 images (basic capability)
- Recommended: 30-50 images (good consistency)
- Optimal: 50-100 images (best results)
Images should show the subject from multiple angles, in different lighting, and with varied backgrounds. Diversity in the training set produces a more versatile LoRA.
Can I combine multiple LoRAs?
Yes. Leonardo supports LoRA stacking—activating multiple LoRAs simultaneously with individual weight controls. A common setup:
- Character LoRA (weight: 0.7-0.9) for subject consistency
- Style LoRA (weight: 0.5-0.8) for aesthetic consistency
- Concept LoRA (weight: 0.3-0.5) for specific visual elements
Higher total LoRA weight can produce artifacts; start with moderate weights and increase gradually.
Can I share or sell my trained LoRAs?
Leonardo’s trained LoRAs are primarily for your own use on the platform. Check the current terms of service for sharing and distribution rights. Unlike Civitai’s open model sharing, Leonardo’s LoRAs are tied to the platform’s generation system.
API Access
Does Leonardo have an API?
Yes. The Leonardo API provides programmatic access to:
- Image generation (text-to-image, image-to-image)
- LoRA management
- Generation history
- Model selection
- Canvas features (limited)
API access is available on the Artisan plan.
What can I build with the API?
Common API applications:
- Automated asset generation: Generate product images, marketing assets, or game art from data
- Custom interfaces: Build application-specific generation UIs
- Batch processing: Generate large volumes of images programmatically
- Pipeline integration: Connect Leonardo to design or content management systems
- E-commerce automation: Generate product lifestyle images from catalog data
Are there API rate limits?
Yes. Rate limits depend on your plan tier:
- Artisan: Higher rate limits suitable for production applications
- Exact limits are documented in the API documentation available to paid subscribers
Commercial Rights
Can I use Leonardo-generated images commercially?
- Free plan: No commercial use rights
- Apprentice plan: Limited commercial rights (check current terms)
- Artisan plan: Full commercial rights for all generated content
What commercial uses are covered?
On the Artisan plan, commercial uses include:
- Advertising and marketing materials
- Client deliverables
- Product packaging
- E-commerce product images
- Social media content for business accounts
- Game assets for commercial games
- Book and publication illustrations
Who owns the generated images?
Users retain rights to content generated on paid plans, subject to Leonardo’s terms of service. The standard arrangement grants users broad usage rights while Leonardo retains certain platform-related rights.
Are there content restrictions?
Yes. Leonardo prohibits generation of:
- Content depicting minors in inappropriate contexts
- Non-consensual intimate imagery
- Content designed for harassment or harm
- Deepfakes of real individuals without consent
- Content violating applicable laws
Safety filters are built into the generation pipeline and cannot be bypassed.
Motion and Video Features
Does Leonardo support video or motion generation?
Leonardo.ai is primarily an image generation platform. As of the current version:
- Motion features: Leonardo has introduced limited motion/animation features (Leonardo Motion) that can add subtle movement to generated images
- Full video generation: Not Leonardo’s primary focus; dedicated video tools (Runway, Kling, Pika) are better suited
What is Leonardo Motion?
Leonardo Motion adds subtle animation to still images—gentle camera movements, parallax effects, and environmental motion (swaying leaves, flowing water). It’s designed to bring generated images to life for social media and web content, not to produce full video sequences.
Can I create animated characters with Leonardo?
Limited animation is possible through Motion features, but Leonardo is not designed for character animation. For character animation needs, tools like Higgsfield (photorealistic) or Viggle AI (character motion transfer) are more appropriate.
Technical Questions
What resolution does Leonardo support?
Resolution options vary by plan:
- Free: Standard resolution (up to 1024×1024 or equivalent)
- Apprentice: Higher resolution options available
- Artisan: Maximum resolution options
Exact resolution limits depend on the model used and current platform capabilities.
What file formats does Leonardo output?
Generated images are available in:
- PNG: Default format, supports transparency
- JPG: Available for smaller file sizes
Can I use ControlNet with Leonardo?
Yes. Leonardo supports ControlNet features including:
- Pose control (guide character positioning)
- Edge detection (maintain composition structure)
- Depth mapping (control spatial relationships)
ControlNet availability may vary by plan tier.
What is the real-time canvas?
Leonardo’s real-time canvas is a spatial generation workspace where you can:
- Generate images directly onto a canvas
- Arrange multiple generated elements
- Inpaint specific areas
- Outpaint to extend compositions
- Layer and compose generated content
The canvas supports a more iterative, compositional workflow than the standard prompt-result cycle.
Account and Billing
Can I cancel my subscription?
Yes. Subscriptions can be canceled at any time through account settings. Access continues until the end of the current billing period.
Do unused tokens roll over?
Tokens do not roll over between billing periods. Unused tokens expire at the end of each billing cycle. Plan your usage accordingly.
Can I buy additional tokens?
Token top-up options are available on paid plans for users who exhaust their monthly allocation. Check the platform for current top-up pricing.
Is there an enterprise plan?
For organizations with specific requirements (dedicated infrastructure, custom integrations, volume pricing), Leonardo offers custom enterprise arrangements. Contact their sales team for details.
References
- Leonardo.ai Official Website. https://leonardo.ai
- Leonardo.ai Documentation. https://docs.leonardo.ai
- Leonardo.ai API Documentation. Available to Artisan plan subscribers.
- Leonardo.ai Terms of Service. https://leonardo.ai/terms
- Hu, E. J., et al. “LoRA: Low-Rank Adaptation of Large Language Models.” ICLR, 2022.
- Zhang, L., et al. “Adding Conditional Control to Text-to-Image Diffusion Models.” ICCV, 2023.
- TechCrunch. “Leonardo.ai Platform Overview.” TechCrunch, 2025.