Leonardo AI
by Leonardo.Ai (Canva)
Generative AI suite for image, art, and video with custom model training
Last reviewed 2026-06-20
Leonardo AI (Leonardo.Ai) is a generative AI platform for creating images, art, and video from text prompts and reference images. It pairs its own foundational models (Phoenix, Lucid Origin, Lucid Realism) with third-party models such as FLUX, Ideogram, GPT Image, Nano Banana, and Veo, all behind one creative web app and a developer API. Beyond text-to-image, it offers image-to-image, image-to-video, an AI Canvas for inpainting and outpainting, a Universal Upscaler, background removal, 3D texture generation, and custom model (LoRA) training, originally aimed at game artists and now used broadly by designers, marketers, and creators. Leonardo is best understood as an assistant: it generates and edits assets on request using sliders and visual controls, and a person selects, refines, and decides what to keep, so it is not an autonomous agent. Founded in Sydney in December 2022, it grew to a large registered-user base (vendor and press figures vary, reported in the tens of millions) and raised a Series A before being acquired by Canva in 2024; it continues to operate as a standalone product and API while feeding Canva's wider AI tooling.
What it can do
Text-to-image and image-to-image generation
AssistantGenerates photorealistic and stylized images from text prompts or reference images using Leonardo's own models (Phoenix, Lucid Origin, Lucid Realism) plus third-party models such as FLUX, Ideogram, GPT Image, and Nano Banana.
sourceImage-to-video and motion generation
AssistantAnimates static images and generates short video clips; the platform integrates motion and video models including Google's Veo 3, Kling, LTX, and Seedance via its app and API.
sourceAI Canvas and image editing
AssistantProvides inpainting and outpainting on an editing canvas, a Universal Upscaler, background removal, and unzoom to refine generated artwork.
sourceCustom model and 3D texture generation
AssistantLets users train custom models (LoRA) on their own datasets and generate contextual textures for 3D assets, supporting consistent styles and game-art workflows.
sourceProduction API for generation workflows
AssistantA developer API exposes image, video, and 3D generation (text-to-image, image-to-image, image-to-video) across in-house and third-party models under a pay-as-you-go model, for embedding generation in other products.
source
Strengths
- +Broad model choice (own Phoenix/Lucid plus FLUX, Ideogram, GPT Image, Veo) in one app
- +Strong creative controls: AI Canvas, upscaling, custom model (LoRA) training, 3D textures
- +Both a self-serve web app and a production API for developers
Limitations
- −An assistant, not an autonomous agent: a person selects and refines every output
- −Token-based pricing can add up for high-volume image and video generation
- −User-count and generation totals are vendor and press reported, not independently audited
Overview
Leonardo AI (Leonardo.Ai) is a generative AI suite for images, art, and video. It combines its own foundational models (Phoenix, Lucid Origin, Lucid Realism) with third-party models (FLUX, Ideogram, GPT Image, Nano Banana, Veo, Kling, and more) inside a single creative web app and a developer API. It started with a focus on game and concept artists and broadened to designers, marketers, and general creators.
What it does
It generates images from text prompts or reference images (text-to-image, image-to-image), animates stills and produces short clips (image-to-video, motion), and edits via an AI Canvas with inpainting and outpainting, a Universal Upscaler, background removal, and unzoom. It supports custom model (LoRA) training on user datasets and contextual 3D texture generation. A person drives generation with sliders and visual controls and selects what to keep, so Leonardo is an assistant rather than an agent.
Integrations & setup
Self-serve via web app and native iOS and Android apps, plus a production REST API for image, video, and 3D generation under pay-as-you-go pricing. Since the 2024 Canva acquisition it also feeds Canva's AI tooling. No public MCP or A2A support is documented as of this review.
Pricing
Freemium. A free tier grants a daily token allowance (reported around 150 tokens/day). Paid plans are token-based and reported at roughly Apprentice $12/month, Artisan $30/month, and Maestro $60/month, with discounts on annual billing; higher tiers add more tokens, custom model slots, and relaxed generations. API usage is billed pay-as-you-go per generation. Confirm current numbers on the pricing page.
Best for / not for
Best for artists, game and concept designers, marketers, and developers who want a wide model selection, custom training, and both an app and an API in one place. Less suited to those wanting a hands-off autonomous creative agent, or pure true-vector output (see Recraft).
Traction
Founded in Sydney in December 2022 by JJ Fiasson and co-founders, it raised a Series A (reported around US$31M, led by Blackbird) before being acquired by Canva in 2024 (deal value reported around US$78.5M by press). Registered-user totals are vendor and press reported and vary by source (tens of millions), so treat exact counts as unverified.
Alternatives
Midjourney and Ideogram focus on image generation quality and typography; Recraft targets brand-consistent design and true vectors; Playground AI is another web-based generator.
What people are saying
We aggregate real LinkedIn discussion into sentiment for the agents people search most. Leonardo AI isn't tracked yet, want it added? Request tracking.
FAQ
Is Leonardo AI an autonomous agent?+
No. It is a generative AI image and video suite. It produces and edits assets on request, and a person selects and refines the output, so it operates at the assistant level. Its API lets it serve as generation infrastructure inside other products.
Who owns Leonardo AI?+
Leonardo.Ai was founded in Sydney in December 2022 and was acquired by Canva in 2024. It continues to run as a standalone product and API while contributing to Canva's broader AI tooling.
What models does Leonardo AI use?+
Its own foundational models (Phoenix, Lucid Origin, Lucid Realism) alongside third-party models such as FLUX, Ideogram, GPT Image, Nano Banana for images and Veo, Kling, LTX, and Seedance for video.
Sources
- Leonardo.Ai homepage · accessed 2026-06-20
- Leonardo.Ai API for developers · accessed 2026-06-20
- Lucid Origin model page · accessed 2026-06-20
- Leonardo.ai (Wikipedia) · accessed 2026-06-20
- Canva acquires genAI startup Leonardo.AI (MarTech) · accessed 2026-06-20
Last reviewed 2026-06-20