
⠋ building cool stuff ⠙ 23 ┃product engineer @HeyGen based in Toronto ┃ Built @editfa_st ┃Building @HyperFrames_ by HeyGen
🇪🇸/🇨🇦/🇺🇸
Joined August 2023
- Tweets13.7K
- Following1.5K
- Followers3.8K
- Likes76.8K
Pinned Tweet
just a happy boy from a small town in Spain, here in SF
thanks @HeyGen_Official @joshua_xu_ @liu8in for making it possible ❤️
WOW
HeyGen Video just hit #1 on @vercel AI Gateway 👀
vercel.com/ai-gateway/leader…
Miguel Ángel retweeted
Great writeup on agentic video which we've been working on quite a bit with @HyperFrames_
We've been able to democratize video creation internally because of agentic editing
and it's how we are able to create launch videos every day without needing a phd in after effects or premiere pro
A few thoughts on agentic video editing I’ve had over the last few days:
1. 99% of ads should be edited agentically now. If your editors are still editing in the timeline, they’re falling behind. I got a message from my senior VE saying that he’s putting out 5x edits/day now at the same quality he was before.
2. The conversation with video editors will be uncomfortable because the timeline is all they’ve ever known. Some will be hesitant and some will be fearful for their job. I think the best way to address this is head on. The tech is inevitable, but having taste is still going to be incredibly valuable to train and give feedback to agents. Those who have this will become more valuable, those who don’t will struggle.
3. Opus 5.5 can natively edit all of your ads without needing any MCPs or plugins. I’ve tried the alternatives and couldn’t find anything that was significantly better than doing everything inside of Claude.
4. Have your senior editor(s) build skills for each style of edit (e.g. mashups, UGC, AI animations…). You can build these by feeding 4-5 inspo edits to Claude and tell it to make a skill out of it. I prefer building lots of mini-skills to one big video editing skill, so that the context stays as relevant as possible.
5. Once you have a base skill (e.g. ugly ads edits), the best way to train it is through feedback loops. Have the skill re-edit the last 25 ads each senior edited then put them in an artifact, and have the senior VE give feedback on each. Keep doing this until the skill is refined and it understands what a good edit is.
6. If you do this, there’s no reason why you shouldn’t be able to make edits that need little/no feedback. Most of ours are approved after v1 or v2, sometimes v3.
7. Only A-player editors should be training the skills.
8. Re adoption in the team, the AI champion builds the v1. Then get out of the way. I’m not the best editor on my team, so my job is to build the harness that lets the best editors train the skills, not build the skills myself.
9. If you want to supercharge this process, have your top editors record their screen as they edit. Turn on Gmeet or loom and get them to narrate the reasoning behind their process. E.g. why they chose this cut, pacing or b-roll clip…this is all context that gets fed into your skills.
10. It takes time to build, train and refine these skills. You have to go through the dip to get out the other side. Budget time in for senior VEs to work on this - otherwise you’ll end up with skills that put out subpar edits and don’t save you much/any time.
11. Operate under the assumption that this is the worst video editing is ever going to be. Opus 5.5 can do most of our editing today…what will Opus 6/Fable 6 be able to do? The editing skills you build are only going to become more valuable as the tech evolves.
I still don’t think people realise how big video editing being unblocked is. Things that I thought were 3-6 months away are now on the table.
Will continue to share more as I learn. CRAZY, crazy times.
Miguel Ángel retweeted
everyone's posting videos made with Opus 5.5, but most don't show you how
Pit Crew Academy is live in HyperFrames Studio
we turned months of making videos and templates into lessons from beginner to pro
anyone can create pro videos now
get started ↓
Miguel Ángel retweeted
"over 1,000 medieval scribes have joined the Association for Human Manuscripts, pledging to never use the printing press or work with Gutenberg
654 have gone even further, promising to never read a printed book"
Miguel Ángel retweeted
SWE-bench tells you which model writes working code
Code2Video tells you which one makes a video people want to watch
it's now a standalone leaderboard on @ArtificialAnlys
27 models, judged head to head on engagement, prompt intent, composition, temporal consistency and craft
Claude Opus 5.5 still leads
Check out their article here ↓
Introducing Code2Video-AA, our independent implementation of HeyGen’s Code2Video Bench, which evaluates how well models create product and brand marketing videos in code
Code2Video-AA tests models across 114 design briefs, tasking them with turning each product or brand marketing brief into a @HyperFrames_ HTML composition, which we render to video. Model outputs are evaluated through pairwise comparisons with other model outputs across five axes: Engagement, Prompt Intent, Composition, Temporal Consistency, and Craft. We use the outcomes of these comparisons to calculate an Elo rating for each axis, which are then averaged to produce the headline Code2Video-AA Elo. Code2Video-AA was built in partnership with @HeyGen.
Claude Opus 5.5 (max with fallback, 1536) leads Code2Video-AA by 93 points, but costs $3.08 per task, ~10x more than GPT-6.1 Sol (max, 1369), which trails by 167 points.
Key findings:
➤ Claude Opus 5.5 leads on all five judge axes: Models are judged across 5 axes: Engagement, Prompt Intent, Composition, Temporal Consistency, and Craft. Claude Sonnet 5.5 (max with fallback, 1443) and GPT-6 Astra (max, 1427) are statistically tied for second in overall score
➤ Cost per task varies more than 300x, from $0.01 to $3.89: GPT-6.1 Sol (max), fourth, has a Cost per Task of $0.31, about 90% lower than Claude Opus 5.5 (max with fallback) at $3.08
➤ Open weights models trail by at least 483 points: DeepSeek V4.1 Flash (max, 1053), MiMo-V2.6-Flash (1052) and MiMo-V2.6-Pro (1052) are the top open weights models, at $0.01 to $0.08 per task
Code2Video-AA details:
➤ Models are tested on a single response without tools. We render it with HyperFrames, with no network access
➤ A judge (Gemini 4 Argon, high) compares pair of videos across five axes: engagement, prompt intent, composition, temporal consistency and craft, with the final score being calculated from a combination of each individual axis Elo
➤ Over 35,000 verdicts rank 27 models by Elo, with 95% confidence intervals
More results, example videos and how Code2Video-AA works follow in the thread below
LFGGG
Introducing Code2Video-AA, our independent implementation of HeyGen’s Code2Video Bench, which evaluates how well models create product and brand marketing videos in code
Code2Video-AA tests models across 114 design briefs, tasking them with turning each product or brand marketing brief into a @HyperFrames_ HTML composition, which we render to video. Model outputs are evaluated through pairwise comparisons with other model outputs across five axes: Engagement, Prompt Intent, Composition, Temporal Consistency, and Craft. We use the outcomes of these comparisons to calculate an Elo rating for each axis, which are then averaged to produce the headline Code2Video-AA Elo. Code2Video-AA was built in partnership with @HeyGen.
Claude Opus 5.5 (max with fallback, 1536) leads Code2Video-AA by 93 points, but costs $3.08 per task, ~10x more than GPT-6.1 Sol (max, 1369), which trails by 167 points.
Key findings:
➤ Claude Opus 5.5 leads on all five judge axes: Models are judged across 5 axes: Engagement, Prompt Intent, Composition, Temporal Consistency, and Craft. Claude Sonnet 5.5 (max with fallback, 1443) and GPT-6 Astra (max, 1427) are statistically tied for second in overall score
➤ Cost per task varies more than 300x, from $0.01 to $3.89: GPT-6.1 Sol (max), fourth, has a Cost per Task of $0.31, about 90% lower than Claude Opus 5.5 (max with fallback) at $3.08
➤ Open weights models trail by at least 483 points: DeepSeek V4.1 Flash (max, 1053), MiMo-V2.6-Flash (1052) and MiMo-V2.6-Pro (1052) are the top open weights models, at $0.01 to $0.08 per task
Code2Video-AA details:
➤ Models are tested on a single response without tools. We render it with HyperFrames, with no network access
➤ A judge (Gemini 4 Argon, high) compares pair of videos across five axes: engagement, prompt intent, composition, temporal consistency and craft, with the final score being calculated from a combination of each individual axis Elo
➤ Over 35,000 verdicts rank 27 models by Elo, with 95% confidence intervals
More results, example videos and how Code2Video-AA works follow in the thread below
meme made with @HyperFrames_, you can just create great memes with it and copy any meme too inside the desktop app btw
Miguel Ángel retweeted
Improved this video with HyperFrames studio and one prompt
give your features the launch they deserve with @HyperFrames_
the best way to understand the S1 filing from @AnthropicAI using @HyperFrames_ Studio!
I'm not reading 300 pages, so I let Opus 5.5 create this video in @HyperFrames_ Studio to explain the 10 most interesting pieces of the S1 filing
Miguel Ángel retweeted
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 👇
kurzgesagt animation style has been totally reversed engineered using @HyperFrames_!
This quoted post is unavailable.
install the desktop app to be able to do it! hyperframes.dev
Miguel Ángel retweeted
opus 5.5 era changed how code-to-video works
so we deleted ~98k lines of skills from @HyperFrames_ 🤯
smart models don’t need a workflow. they need the best tool to make video
21 skills → 7
470 KB of instructions → 39 KB
and it got way better:
→ agent uses the tools right: 13% → 79%
→ fully correct videos: 11% → 71%
→ 68% cheaper, 76% faster per run
we had to remove 100k LOC of skills in the @HyperFrames_ repo thanks to Opus 5.5
now we have only 7 skills instead of +20 with better quality and speed (from init to finished video) thanks to the herculean efforts from @wenbozhu0328 and @WaterrrMiao!
Miguel Ángel retweeted
Here’s how to use Claude to copy any video shorts style in a few minutes.
From @jake11moran:
“We added this as a native capability in @HyperFrames_ studio. You just upload or pick a reference clip and then describe in the prompt what you want to change.
You can see where we’re drawing inspiration from and apply a different font or different colors. It’s still structurally the same idea.”
📌 Watch our full AI video editing tutorial here: youtu.be/MC1g2fHYC88
Anyone can now make amazing videos with Claude Opus if you know how to prompt it.
Here's my new episode with @jake11moran where he showed me exactly how to use Claude and @HyperFrames_ to:
→ Make a launch video from just a few images
→ Duplicate your favorite creator's video style
→ Make precise edits with the right keywords
Every minute of our 40-min episode is packed with alpha and live demos.
📌 Watch now: youtu.be/MC1g2fHYC88
Thanks to our sponsor:
@WisprFlow: Get 1 month free of the best voice dictation tool wisprflow.ai/?promo_code=PET…
Best motion design library for Claude with +300 free video components
hyperframes.heygen.com/catal…


























