Miao Yang
@WaterrrMiao • 2,123 subscribers
product engineer @HeyGen based in Singapore | building the next generation of agents @HeyGen | cooking fancy videos with @HyperFrames_
Videos
Kurzgesagt has 25M subs. Millions of views on EVERY single video 🤯 nobody copies them for a reason: one video takes 1,200 hours and a studio of 50+ animators. we rebuilt their format with Opus 5.5 + HyperFrames Studio. 100% code. a full 11-minute explainer. made in ~2 hours. under $50. want the template? we’re open-sourcing it for free 👇
Miao Yang353,606 views • 18 days ago
复刻小Lin说 1 分钟，现在谁都能做。 做满 10 分钟、还是那个味道，是另一回事🤥 Opus 5.5 之后，code to video 到了新的高度。 复刻一个片段 → 完整复刻 10 分钟 → 创作新的 10 分钟。 HyperFrames editor 已经在第三步了：「一口气了解泰国经济」，9 分 34 秒，一集她没讲过的题目。 最近 reverse engineering 很火，刷屏的小Lin说复刻帖我都看了。大多数复刻的是一段 1–2 分钟的 motion graphics：形似，但简单、廉价。小Lin的动效细节是满的，动作、编排、前后的连续性，全都和口播文案咬在一起。这不是复刻一分钟的片子能做出来的。 素材可以照着复刻。但如果从 0 开始呢？如果要一集接一集地做呢？ - 素材去哪找 - 每句话配什么画面，怎么卡着文案进出 - 文案怎么写才是小Lin味，而不是别的博主，或者流水线式的讲解 - 数字人和声音怎么撑满 10 分钟：语序、语调、停顿都要对 - 音效怎么提取、归类，用在对的地方 - 背景音乐在哪起、在哪停 到这一步，agent 就一头雾水了。 HyperFrames editor 把这些都解决了： - 把她拆成一套模板：怎么写、怎么剪、怎么动 - 动效逐帧对齐，1:1 重建 - 音效和音乐跟着口播卡点 - HeyGen 数字人 + 专业级声音克隆 下面泰国这集就是这么来的：同一套模板，换个题目就是新的一集。 HyperFrames editor 做的不是复刻小Lin说，是一个能无限往下做的「一口气了解 XXX」系列。 （视频中的主持人形象和声音均为 AI 生成，非小Lin本人，仅作技术演示）
Miao Yang462,894 views • 25 days ago
we one-shotted a 7-minute true-crime documentary in HyperFrames Studio 🎬 not a clip. a full film: the $4.5B Bitfinex heist, made in the style of Fern's investigations. script, voice, edit, sound, footage, 3D. HyperFrames studio did every part. what makes 7 minutes hard 👇 ✂️ the edit 37 scenes, 10,525 frames, every cut timed to the narration. the agent measured Fern's pacing and how their films split between 3D, diagrams, documents and footage, then cut all 7 minutes to match. 🔊 the sound 109 sound effects, each timed to the frame. music comes from HeyGen's audio library: the agent auditioned 41 tracks against Fern's soundtracks, kept the closest, and hard-stopped them on the big reveals. 🎙️ the voice the whole narration is generated with HeyGen Voice 1. the agent transcribes every line back and re-records any it misreads. 🗂️ the footage the agent sourced everything itself with our skills: court filings, DOJ releases, Bitfinex's 2016 blog via the Wayback Machine, the blockchain record of the stolen coins, real logos. then it cropped, graded and animated every piece. what it couldn't find, it built: almost 4 minutes of 3D in HyperFrames . this format belongs to Fern, some of the best investigative storytelling on YouTube. all credit to their team. go watch the originals 🙏
Miao Yang79,437 views • 14 days ago
Every video you see is now a template. Top: OpenAI's animated brand film. Bottom: /reverse-any-video. One prompt, no human edits → all 13 cuts on the exact frame → every shot re-made to the original's motion → every word rebuilt as editable HyperFrames code Reverse is step one. Then remix it: new script, new brand, your product. Works on anything you scroll past: • launch videos + product demos • ads + UGC • explainers + tutorials • talking heads + podcast clips • UI animations + motion graphics • animated shorts + brand films • 3D camera moves • maps + data viz • kinetic type + lyric videos • trailers + YouTube intros • AI-generated clips How to use it: 1. npx hyperframes skills update 2. Open Claude Code (or any agent with skills) 3. /reverse-any-video + paste a link or file It watches frame by frame, rebuilds every layer as a HyperFrames project, then checks its render against the original. What would you reverse first? 👇
Miao Yang43,013 views • 9 days ago
OMG. Fable 5.2 is sooooooo crazy. I tried it with HyperFrames , using Reference to Video in HyperFrames Studio -> The HyperFrames team shipped way too many features in the last 2 weeks, so I picked 13 of them and one-shotted a launch video for each: 1. Storyboard 2. Fork a chat 3. Voices 4. Comments on the preview 5. Podcast edit 6. HeyGen Video 1 7. Projects already on Home 8. OpenClaw + Hermes Agent 9. Remove background 10. Search your skills 11. Your own shortcut keys 12. HeyGen Lipsync 13. Brand kits How it works: -> drop in any video you love -> pick what to copy -> get your own version back. 1/ Storyboard is right here(ref: Framer ). The other 12 are in the thread 🧵:
Miao Yang10,615 views • 3 days ago
We tested a lot on this: 20 films, simple to complex, from UGC to talking heads to motion graphics. Our reverse-any-video vs REA-Video, two experiments: 1. Analysis: cuts, editing, how each shot was made 2. Full 1:1 remake: same model, same keys, isolated containers Experiment 1, analysis (hand-labelled cuts), ours vs REA-Video: • Cuts found: 108/108 vs 59/108 • Precision: 96% vs 89% • Time: 81 s vs 1,595 s, ~20× faster • Its misses: talking heads, heavy edits • It called live footage AI in 3 of 8 Experiment 2, full remake, ours vs REA-Video: • SSIM: 0.96 vs 0.78 Bottom line: • Simple UGC: both great • Talking heads with motion graphics, B-roll and heavy editing, and launch videos with complex motion graphics: we do better • REA-Video remakes slightly faster and cheaper example:
Miao Yang20,676 views • 7 days ago
No more content to load