Build video with code
New York, NY
Joined October 2017
- Tweets53
- Following2
- Followers1.3K
- Likes1.1K
Editframe retweeted
Is it time to build the best damn video editor the web has seen?
Local first, multiplayer, offline support, figma levels of ux, claude design levels of agentic control.
Powered by @ElevenLabs, @editframe, mediabunny, mobx, plus a few homegrown libs
Editframe retweeted
Day 4: Building an Escape Room entirely using AI (GPT-Astra). and the tools to design it.
Level 2 is taking shape.
Three.js + Blender + React Flow
Video made using @editframe agents
Models partially made using @MeshyAI
#gamedev #threejs
Editframe retweeted
Building a video pipeline with @editframe to generate some killer wildcard recap videos (Opus 5.5 is beastly here)
Editframe retweeted
i had fable recreate this launch video.
now imagine connecting your @figma mcp, and giving your agent access to your customer list, a database, and rendering a personalized video for each of them for this launch.
1000's of renders, in parallel in ~1 min.
swap out text, media, designs, genai, change aspect ratios, anything really.
everything is deterministic.
editframe.com/cloud
Editframe retweeted
ok ok here me out.
lets recreate this @QuiverAI video using Astra and @editframe
🤯 but let’s use Arrow 2 to make the svg’s in the video 🔥🔥
i didn’t think this would work so well 😂
Editframe retweeted
pace deez videos
@EricSimons open weights are cool, but what about remaking a video using @editframe?
every agent/model is already trained on Editframe, it’s just html/css
adding this to the repo later today, take it for a spin, github👇
Editframe retweeted
switched back to opus for this one, and still pretty blown away recreating this @ElevenLabs video with @editframe
👀 see if you can spot the difference at the end
sound on 🔊
took a bit of nudging, but the agent got really really close. take this for a spin, github repo is in the next post 👇
Editframe retweeted
I tried to single prompt-shot your motion design using GPT-6 Astra + @editframe.
Turned to out be great. Solid 8 / 10.
Yours is ofcourse 10 / 00
@kirtanlb
@yudDIDit
#gpt6 #motiondesign
Editframe retweeted
Not bad for a single prompt using Astra and @editframe
I tried to single prompt-shot your motion design using GPT-6 Astra + @editframe.
Turned to out be great. Solid 8 / 10.
Yours is ofcourse 10 / 00
@kirtanlb
@yudDIDit
#gpt6 #motiondesign
Editframe retweeted
👀
TsugiteがEditframe(@editframe )に対応しました 🎬
公式サンプル27種を、ブワーッと広げてみた。
ひとつひとつ、中の映像も動いています。
この告知動画も、Editframe公式サンプルをアレンジして制作。
つくりたい表現から選べるの、楽しい。
制作の仕組みはツリーに👇
Editframe retweeted
Editframe（@editframe）の公式「Claude Cowork」テンプレートをベースに、セミナーの告知動画を作ってみました。
Codexと相談しながら、構成や動きを調整。
背景・配色・書体は僕のWebサイトに合わせ、BGMと効果音はElevenLabsで別途制作しています。
そして、動画でお知らせしているセミナーがこちら👇
AIに見つけてもらうWebサイトへ
〜AEO対策とエージェンティックインターネット入門〜
「おすすめのお店を教えて」
そうAIに聞いたとき、あなたのお店やサービスは見つけてもらえるでしょうか。
AIに情報を見つけてもらい、正しく理解してもらうには？
さらに、問い合わせや予約などの行動につなげるには？
「自分のサイトなら、何から始めればいい？」
そんな疑問を、身近な例と実際の取り組みを交えながら、一緒に整理する60分です。
専門知識やプログラミングの経験は不要です。
お店や会社、ご自身の活動のWebサイトがある方も、これから作りたい方も、お気軽にどうぞ。
詳細はリプから👇
Editframe retweeted
shared with Astra this @bot video and asked it to use @editframe to recreate it.
👀 spot the difference?
will add this to our GitHub repo later today👇
Grok Bot can now buy things on the internet.
To get started, connect @link, send your Bot shopping, and it can complete purchases on your behalf.
Editframe retweeted
I'm still looking but:
- Got a couple really cracked editors that charge a decent rate
- Got some editors which aren't that good who charge 3-4x more then those really cracked editors
At this point it might be easier just to use GPT 6 Astra + @editframe 👀
Editframe retweeted
Dropped this storyboard into Codex with Astra and told it to use @editframe to turn it into a video.
Editframe retweeted
Everyone wants to know 👀👀
First off, Google Gemini is pretty incredible when it comes to motion, sequencing, video understanding, etc.
training on YouTube probably helps :)
All of the models are really good at Editframe, our abstraction has an edge and unlocks capability.
Feel free to test or benchmark, but outcomes matter when you’re burning tokens 🔥
For most of these videos, I was using a recent sonnet or opus model to drive the agent. I would have the agent upload the video, analysis was being done via Gemini, that info/semantics was being fed back to the coding model, and so on.
At that point, depending on how you prompted, it would compare and adjust.
The things that matter are the semantics of what is happening, the LLM native framework (html/css, react), the temporal model (Editframe), and compute execution layer (Editframe) for feedback (how quickly something renders)
Tbh I think anyone could get these results with not much prompting or video experience. The edge is @editframe
Editframe retweeted
recreated the first 20s of Figma Motion launch using @editframe, took a few prompts.
did not use Astra to make this 🚨, or any of the videos I’ve shared recently.
existing LLMs are already very good at html/css scripting, and have been capable for many many months.
Editframe retweeted
🧵all of these videos were made in a few prompts using @editframe sound on 🔊
- no screen recordings
- just html/css, and scripting
- github repo👇
open sourced as templates
please fork, use these video, agents need video
Editframe retweeted
Just experimented with @editframe - by @yudDIDit to automate some video generation using code, and the dev experience is cool. Managed to pull together this video in record time.
If you're interested, check out the PR here: github.com/editframe/example…
#buildinpublic #editframe
Editframe retweeted
this sounds fun! we're solving some of the hardest problems in video @editframe (HTML/CSS, works with React)
• True browser rendering (DOM/Canvas) editframe.com/sdk
• GUI components that play nice editframe.com/editor
• Cloud editframe.com/cloud