Start from any site
Every layer, colour and font comes in editable - colours become variables, type becomes styles, all bound.
The motion canvas for Claude Code
Paste any website. Claude rebuilds it as editable layers, animates it on a real timeline, and you ship it as video or code.
or run it locally: npx feega
Open source.
Apache-2.0.
Works with Claude Code,
Codex, Cursor.
Graphic & motion design
Hundreds of ready-to-use templates
Made for humans and agents
Ready for Claude & Codex, even from your phone.
Why not Remotion?
Remotion is code: every tweak is a diff and a re-render.
Feega is a canvas: you see it live and drag it.
No lock-ins
Switch between different providers
Bring your own models: OpenRouter, ElevenLabs and more.
Warning - Only for real nerds
Ship it as video or code.
MP4 in every format, or React, Svelte and HTML.
Free to run locally.
Pay only for speed in the cloud.
Local version
Features
- Design + Motion, infinite canvas
- Text, image, video, audio, 3D generations
- Bring your own keys for AI generations
- Unlimited canvases & projects
- Design & Motion templates library
- Copy & paste from Figma
- Coder mode for code export
- Embed mode for interactive scenes
Just run npx feega
Pricing
FreeCloud version
Pricing for AI usage (chatbot, AI generations)
100 credits = €1
e.g. €8/mo = 800 credits included (100-120 pictures)
Everything in local, plus:
- All models from OpenRouter, Kie, Higgsfield, Wiro, Elevenlabs
- Unlimited canvases & projects
- AI assistant in-editor
- Use with your agents via MCP
- Unlimited team members
- Real-time collaboration
- Brand & products import
- Social accounts & posts import
- AI influencers
- Embed mode for interactive scenes
Pricing
18 / moLet your agents do the work
claude mcp add --transport http feega https://mcp.feega.app/mcpBuilt by a designer, for designers.
Feega didn’t start in a boardroom. It started with one designer, tired of paying for slow tools that were never built for working with AI. No VC money. No roadmap written for investors. Just a tool built by a designer, for designers, to make everyday work faster. And it’s open source, so it stays yours.
— Andrea Buttarelli
FAQ
What is Feega?
Feega is one infinite canvas for design, motion and AI generation. It is open source: use it in the browser, or run it locally with npx feega. Google Sign-In is used only for authentication; no Google API is used to generate or edit content.
Is Feega open source?
Yes. Feega is open source under the Apache-2.0 license. The code is on GitHub.
Can I run it locally?
Yes. Run npx feega: no account, no cloud. Your work is saved in one .feega file on your machine. You need Node 22.
How does the canvas work?
Design in frames with layers and components, animate them on a motion timeline, and add AI nodes for text, images, video, 3D and audio. Wire nodes together so one output feeds the next, then run them to generate.
Which AI models can I use?
Images: Nano Banana (Google), GPT Image (OpenAI), Seedream (ByteDance) and Qwen. Video: Seedance (ByteDance), Kling and Grok Imagine (xAI). 3D: Trellis (Microsoft) and Hunyuan3D (Tencent). Voice, sound and music: ElevenLabs. You pick the model for each node.
Can I use my own AI keys?
Yes, in local mode. Add your keys for OpenRouter, Kie.ai, Higgsfield, Wiro, ElevenLabs and a text LLM provider, and generate from text with them. No plan or credits needed.
What file formats can I import and export?
Import images, video, SVG and Lottie files, paste from Figma, or open a .feega file. Export stills as PNG, JPG, WebP, SVG or PDF, and motion as MP4, GIF, APNG or Lottie.
Can I bring my Figma designs?
Yes. Paste from Figma and your layers stay editable on the canvas.
Can I collaborate with my team in real time?
Yes. Invite teammates to your workspace as members or admins: everyone works on the same canvas and sees changes and who is online live.
Can AI agents use Feega?
Yes. Agents like Claude or Codex can work on your canvas through our MCP server and REST API, acting on your behalf with your permissions.
Are generated files marked as AI-generated?
Yes. Every image and video Feega renders carries a machine-readable AI-generated marker in the file, and when we publish for you we set the platform’s AI-content label where one exists. See AI Transparency.
Who owns what I create?
You keep ownership of the content you bring, and you may use what you generate for your own business purposes, subject to our Terms and third-party rights.
Is my content used to train AI?
No. We do not use your prompts, uploads or generated content to train AI models. See our Privacy Policy.
How is content kept safe?
Feega must never be used to create non-consensual intimate imagery (NCII), sexual deepfakes or any sexual content of real people. Sexual content involving minors is forbidden and reported to the authorities. Prompts and reference images are screened automatically, real or identifiable people are blocked in sensitive contexts, and AI outputs are labelled as AI-generated. Anyone can report content to support@feega.app: we remove it and suspend the accounts responsible. Read our Acceptable Use Policy, Terms and AI Transparency.
How do credits and pricing work?
In the cloud, AI runs are paid in credits. Every €1 of your plan gives 100 credits (€8/mo = 800 credits). A run costs 120 credits for every $1 the AI provider charges. Running locally with your own keys needs no credits. See pricing.
Who is behind Feega?
Feega is built in Italy by Andrea Buttarelli, a freelance designer who wanted a faster way to work. No VC money, no growth team: just a designer building the tool he needed, in the open. Company details are in the Legal Notice, or write to hi@feega.app.