Every time I created new products my focus always went first (wrongly) to building instead to distributing and helping others to use what I created.
After building my thing and having all the excitement of seeing the new toy working, the promotional and communication part always felt more like an obligation, and most of the times was neglected. I always struggled to create manuals,
videos, support content, etc.
So I open sourced a skill for Claude Code that helps in that last part. With some instructions and this skill, you can create product demo videos. With professional voice overs, and subtitles in different languages. Combining the browser and the terminal. Everything controlled from your AI. For example, the clip in this post was made with it.
It's not going to do distribution for you, but it's a start ;)
Good idea, but the demo video needs some work. The music is too loud and distracting especially with that 4 on the floor kick drum, and it sounds like there's some aggressive sidechain compression to make room for the narration.
> "the promotional and communication part always felt more like an obligation"
Yes, this feels like the worst part for me as a software engineer. It doesn't help that many communities have rules against self promotion. If I spent N hours building a tool for myself[0], I don't want to spend another Nx hours where x is some multiple to share it with others while dodging self-promotion accusations. Luckily HN is much more lenient about this than reddit communities or lobsters, which are also great in their own ways.
That's great, really. i do many small tools for the enterprise, and i know first hand that you cannot assume they will read the docs or know where to click.
This is the sort of skill which is worth building specifically for your corporate style and under your sales team's guidance.
I have one[1] for my demos, which uses computer use + ffmpeg used the way I like it to be.
Whenever my team goes out to a conference or other sales event, they carry with them a "booth" package [2] made out of the demos. The companies and use-cases are pulled the weekend before.
Thanks for sharing, but did you even watch your example output[0]? The voiceover is cut off, the drums are out of sync with the rest of the music (rushing).
I'm not using this specifically (I'm targeting native apps, not web apps, so Playwright is irrelevant to my stack), but I'm building out (as part of a pre-existing and unrelated project) something very similar (for that project) - having Claude write the script that steps through product usage and captures a recording of it.
Having product videos on your marketing site that are missing everything new that you built in the last six months because, what, you can't be bothered to manually record them for every release, you wouldn't over-engineer things by writing a script for it because you have more important things to build... seems like a very 2025, pre-AI opinion.
Does it edit the final video w/ raw FFmpeg with captions/voiceover/cropping/panning etc or just do the playwright part? ive been using the filmcraft cli lately to edit videos w/ claude code, but maybe there is a simpler way? Here is one of my recent edits https://www.instagram.com/p/DfJrVhwMSLP/
That's kinda the point I was trying to show. No disrespect to the author of the skill, but I think it's always worth evaluating what something looks like without a skill, since models these days are really good at doing stuff on their own.
Opus 5.5 and Astra are already very good (because of being RL trained at it) most of these. The way I see skills useful are: 1)If they are created to teach how to use a custom and new CLI or MCP 2) They describe gotchas, tips and tricks that are hard to find using simple search or potentially save some tokens for easier to find ones. The OpenAI and Anthropic teams already said to reduce the skills otherwise as they can also waste context space unnecessarily.
In some cases, the point of a skill is to handle all context to a model, so that they don't need to look up on their own (sometimes filling up the context window with irrelevant stuff in their search)
I currently use Claude Code to create my product videos, and one cool thing is that you can have it recreate the UI instead of recording yourself using it or stitching screenshots. That way the whole UI becomes animatable, and you can change it however you need to fit the video.
I can see why this could be interesting for non-tech people, but c'mon, it's just words added to your context, not something magical. I've seen massive hype over Pockock's skills and took a deep dive into them just to find out it's basically "don't make mistakes" on steroids.
I mean, input tokens are cheap, and people would be so much better with just using speech-to-text instead of all of this crap.
The goal is re-usability and (somewhat) reproducibility. Why would I want to repeat the same input tokens from scratch every time I need to redo the same process? You can store the prompt in your notes, a text file on your desktop, or ... in a skill.
Though I agree that it's often rather overhyped for what it is and flows like the one here would likely need additional fine-tuning to match the user / company needs.
I highly doubt this skill adds much. Even prior to Opus 5.5 these models were good enough to generate videos, look at the demos on https://fretta.ai for example.
Adding audio and sfx is easy too if you add ElevenLabs to the mix.
I suffer from the same problem of marketing for my own open source tool that I built. So now I am attempting to use AI to help me market it. But instead of creating a fully automated harness that generates content on its own and posts it to various places, and gets flagged as slop AI content, it helps me identify the places to post by finding relevant discussions in relevant forums, places. Then I draft my own message.
In fact it found me this thread here, because I am also working on a tool (for AI films in this case, not product videos but still adjacent).
You can ask Claude (or Codex) to create yourself a similar thing. Mine wakes up every morning and finds me places to post, tracks it outcomes etc. Gives me a Kanban board as well to keep track of progress.
Do you mean my marketing tool? If so it is just generated by Claude (Opus 5.5)and specific to me. So you can simply prompt it to generate one for you as well. If you mean the app that I built, it is this one: https://gorenku.com/
46 comments
[ 12.6 ms ] story [ 328 ms ] threadAfter building my thing and having all the excitement of seeing the new toy working, the promotional and communication part always felt more like an obligation, and most of the times was neglected. I always struggled to create manuals, videos, support content, etc.
So I open sourced a skill for Claude Code that helps in that last part. With some instructions and this skill, you can create product demo videos. With professional voice overs, and subtitles in different languages. Combining the browser and the terminal. Everything controlled from your AI. For example, the clip in this post was made with it.
It's not going to do distribution for you, but it's a start ;)
https://github.com/aaitor/skill-video-studio
Yes, this feels like the worst part for me as a software engineer. It doesn't help that many communities have rules against self promotion. If I spent N hours building a tool for myself[0], I don't want to spend another Nx hours where x is some multiple to share it with others while dodging self-promotion accusations. Luckily HN is much more lenient about this than reddit communities or lobsters, which are also great in their own ways.
[0] https://github.com/aleyan/dela
I have one[1] for my demos, which uses computer use + ffmpeg used the way I like it to be.
Whenever my team goes out to a conference or other sales event, they carry with them a "booth" package [2] made out of the demos. The companies and use-cases are pulled the weekend before.
[1] - https://gist.github.com/5d4cb99549f7e81a1433a4a13b20622c
[2] - https://demos.isotopes.ai/pages/chicago-booth/
[0]https://www.youtube.com/watch?v=H6npd3uuqls
Having product videos on your marketing site that are missing everything new that you built in the last six months because, what, you can't be bothered to manually record them for every release, you wouldn't over-engineer things by writing a script for it because you have more important things to build... seems like a very 2025, pre-AI opinion.
https://github.com/storytold/filmcraft
Without the skill: https://non.io/video/diffui-product-vid-opus-only.mp4
With the skill: https://non.io/video/diffui-product-vid-with-skill.mp4
Also a disclaimer that there are several major playwright CVEs, so be careful with this / patch the installs you do have.
In some cases, the point of a skill is to handle all context to a model, so that they don't need to look up on their own (sometimes filling up the context window with irrelevant stuff in their search)
Skill-video-studio: 334k
Vanilla Opus: 203k
Cache read:
SVS: 29M
VO: 17M
Output:
SVS: 152k
VO: 93k
Overall vanilla seems to use less, not more, but that's with a sample size of 1.
I can see why this could be interesting for non-tech people, but c'mon, it's just words added to your context, not something magical. I've seen massive hype over Pockock's skills and took a deep dive into them just to find out it's basically "don't make mistakes" on steroids.
I mean, input tokens are cheap, and people would be so much better with just using speech-to-text instead of all of this crap.
Though I agree that it's often rather overhyped for what it is and flows like the one here would likely need additional fine-tuning to match the user / company needs.
https://www.tiktok.com/@everydaylentils
It has an AI avatar on-screen for the demos (any video with a banner uses the plugin). You just use your own fal.ai key and I spend about $2 per demo https://github.com/mrieck/demoday-claude-plugin#top
Adding audio and sfx is easy too if you add ElevenLabs to the mix.