Today we are opening flare.design in Public Beta.
This is also the first post on the flare.design blog, so it feels right to start with an invitation. If you are using AI to make images, videos, posters, product visuals, social content, or if you are connecting agents like Codex, Claude, and ChatGPT into your creative workflow, we would love for you to try it.
flare.design is not meant to be another prompt box that only gives you one generated image. We are building an AI-native motion canvas: a workspace where AI can help generate, arrange, revise, animate, and export visual work, while creators can still see the structure, make precise edits, and keep shaping the final result.
Why we built flare.design
AI generation models have become fast and powerful. But in a real creative workflow, generation is usually only the first step.
After an image is generated, you may still need to add text, change the format, create product scenes, build social variations, annotate a revision, or turn the result into a short video. Too often, that work bounces between tools: one for generation, one for layout, another for motion, another for export.
We want flare.design to bring those steps back into one workspace. Assets, canvas editing, AI generation, motion, templates, collaboration, and export should all happen around the same editable document.
What you can do in Public Beta
The current beta already supports a complete visual creation path:
- Organize images, videos, text, shapes, frames, and layer structures on a canvas.
- Generate or revise visual assets with AI, then place the results directly back into your project.
- Start from existing media and continue with layout, composition, cropping, annotation, and visual exploration.
- Add lightweight motion, preview the timeline, and export work for short-form or social delivery.
- Use Recipes and community apps to turn repeatable creative workflows into runnable AI apps.
- Connect MCP-compatible agents such as Codex and Claude Code so they can read project context, place media, edit the canvas, or run controlled creative tasks.
For example: generate a product shot with AI, drop it onto the canvas and add a caption, give it a little motion, then export a few-second clip — all without leaving the same workspace.
These capabilities will keep changing. Public Beta doesn’t mean every rough edge is sanded down. It means we want real creators inside the workflow early, helping us understand what feels natural, what breaks down, and what should be rebuilt.
Three things we want to learn
If you try flare.design, we are especially interested in three kinds of feedback.
First, what are you trying to make? Product images, ads, short videos, avatars, portfolio pieces, launch graphics, or a workflow we have not imagined yet?
Second, where does the workflow slow down or fall apart? From generation to editing, editing to motion, motion to export, or from a local agent to the cloud canvas, every bit of friction matters.
Third, which parts should become more automated, and which decisions need to stay in human hands? The core question for flare.design is not whether AI can do everything for you. It is when AI should act, and when it should hand control back to the creator.
Start here
You can open app.flare.design and start creating now. You can also read the flare.design Docs for the basics of canvas editing, AI generation, motion, export, and MCP.
During Public Beta, Discord is our main feedback channel. Join the flare.design Discord to report bugs, request workflows, share what you make, or tell us where the product still feels rough. You can also follow us on X, where we will share product updates, examples, and future release notes.
flare.design is still early. That is exactly why we want it to grow around the way real creators work.
Thanks for trying it.