Pinned Tweet
Day 1 is in motion: Lottie AI is live. 🚀
Describe an animation. Get a real Lottie file you can refine with a follow-up, read as code and drop straight into your product.
Start with 10 free generation.
We’re live now on YouTube!
Join us here: lottie.link/behindthebuild
217
What a week! Join our first live demo that’ll take you behind the build of Lottie AI and Creator MCP.
You’ll see real prompts and real decisions, and you can ask the team anything.
Plus a live challenge: animate your country 🌎.
RSVP now!
RSVP here: lottie.link/behindthebuild
245
RSVP here: lottie.link/behindthebuild
116
Day 4 is in motion: @IconScout API brings millions of design assets into your product.
Your users can search, convert and generate without leaving your app.
Get your API key: iconscout.link/api-tw
Stay tuned all launch week: lottie.link/launch-week-tw
5
529
Day 3 is in motion: Meet Generative Icons.
Your agent can now create animated UI icons for your app using Creator MCP.
It also comes with a library of free, open source animated icons.
Explore Generative Icons: lottie.link/generativeicons-…
Stay tuned all launch week: lottie.link/launch-week-tw
4
602
LottieFiles retweeted
I built a workflow that lets your agent make retro ads for any brand. Just give it a website.
These were made by Fable 5.2 on Creator MCP, the headless motion canvas for agents we just launched.
Concepts for brands I love below 🧵
Reply with your favorite and I'll send you the workflow.
Day 2 is in motion: Creator MCP is live.
The motion engine for agents. Fast and headless.
Run it from where your context lives.
Use your own agent in Claude, Cursor or Codex.
Set it up in minutes: lottie.link/creatormcp-tw
Stay tuned all launch week: lottie.link/launch-week-tw
🔗 Live now: lottie.link/lottieai-tw
🔗 Stay tuned all launch week: lottie.link/launch-week-tw
holy f*ck, opus 5.5 is getting dangerously close to replacing motion designers...
the problem is the video comes out looking like random animations stacked on top of each other.
these 6 repos turn claude into a full motion studio:
1. cinetic [github.com/leonxlnx/cinetic] -> makes claude direct a short film instead of stacking random animations. launch films, teasers, logo stings, ui walkthroughs.
2. earshot [github.com/leonxlnx/earshot] -> gives claude ears. 252 sound effects stored as code, claude listens to its own video and fixes the sync.
3. hyperframes (by heygen) [github.com/heygen-com/hyperf…] -> write html, get mp4. ~450 ready blocks: transitions, titles, charts, 3d. apache 2.0
4. remotion-dev/skills [github.com/remotion-dev/skil…] -> 13 official remotion skills. captions, render, maps, multimedia.
5. claude-motion-design [github.com/howseen-ai/claude…] -> html + playwright + ffmpeg. remake mode: 1:1 copy of any launch video, rebuilt for your brand
6. lottiefiles motion-design [github.com/lottiefiles/motio…] -> teaches claude timing, easing and choreography before it writes a single line of code
lottie plans the timing -> cinetic directs -> hyperframes or remotion renders -> earshot listens and fixes the sound
the trick isn't a smarter model. it's giving claude a director's brain, a sense of timing and a pair of ears.
no after effects. no stock sfx library. no motion designer on retainer. just opus 5.5, 6 repos and a loop.
bookmark this before you open after effects again.
the full pipeline is in the article below ↓
LottieFiles retweeted
AI made the first version of motion cheap. It didn't make the second version easy.
You prompt, you get a clip. Fast, often good. Then the second ask comes in. Change the logo. Slow the ease. Match the brand. And the clip can't move. You regenerate and hope.
That's why so many teams are building motion editors right now. The clip was never the hard part. The second edit is.
We think about motion as three jobs.
Video. Footage, people, light. Generation is great here. You pick the best take.
Product motion. Loading states, onboarding, a button that reacts. Tiny, sharp, live on every platform. It changes every sprint.
Launch and explainer video. Built from scenes, rendered at the end. Tomorrow someone still needs one word fixed in one frame.
The last two need a scene that stays editable. Layers, paths, timing, easing. Restyle one shape and leave the rest. Swap an asset and keep the motion.
Video tools will keep getting better at the take. Editable tools will get better at the edit. Most teams need both.
So the question we ask of any motion tool is simple. After the first prompt, can you still make one small change?
That's where craft lives. Not in the first result, but in every edit after it.
Code Meets Animation is coming to Singapore 🇸🇬
It’s a hands-on afternoon of talks, a live Lottie AI Challenge with prizes and some of the best AI builders in town.
🗓 Thursday, 12 November
⏰ 14:00 - 17:00 GMT+8
📍 Lorong AI @ One-North
No animation experience needed. Just bring your laptop.
👉 Register on Luma: luma.com/9dr4cuov






