mamitsu
An API and MCP that watches social videos and returns JSON
34 followers
An API and MCP that watches social videos and returns JSON
34 followers
mamitsu is an API and MCP server that turns the content of a social media post into JSON. It watches the video, listens to the audio and reads the caption and post text, so the ingredients, products and steps in a post come back as data.




Free Options
Launch Team

AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted
Hi Product Hunt 👋 I'm Unai, and I'm building mamitsu.
I kept saving reels of places and recipes and absolutely anything, but never ever revisited them. I built mamitsu initially as part of a more niche app, and then decided to roll it out on its own! Since then, I manage my recipe library and trip spot ideas with my AI agents, and even built a couple apps for some more elaborate integrations.
In short: mamitsu watches a social media post the way a person does. It looks at the frames, listens to the audio and reads the caption, then returns the data you ask for as JSON. Send it an Instagram reel of a trip and get the places with times. Send it a TikTok recipe and get the ingredients with quantities.
You can use it two ways:
→ Build an integration that uses the API, with one POST request, and pass a JSON Schema if you want an exact shape. Think recipe apps, scripts that verify that your brand is actually shown and promoted in influencer videos, etc.
→ From your AI agent, through the MCP server. Connect it and your agent can watch reels, TikToks, YouTube videos and X posts. Ask it to extract tutorial steps, add a recipe's ingredients to your shopping list, anything you can think of.
New accounts get 200 free credits during early access, and credit purchases are doubled while the launch offer lasts. A credit covers about a minute of video (it's all explained in the pricing page).
I'd love to know: what would you build if your app or agent could understand any post? Reply here and I'll run your idea through mamitsu.
https://usemamitsu.com
I found mamitsu to be a really good tool and I’m integrating it into my coffee project, previously I was relying on only screenshots and pictures, but now I can integrate it directly into my app’s user flow; so far the API, documentation and overall experience has been great! Great product @uesteibar !
@jepser Thank you Jepser! Looking forward to seeing that integration ^^ let me know if I can be of any help 🫶
I use it to extract summaries of all the youtube videos I consume. It works great!!
Congrats Unai!
@solidjuls thanks Juli!
I used it, and it's great! It's really convenient to extract structured info from Youtube videos or Instagram reels.
Congrats on the launch, Unai 😊.
@marciobarrios thanks Marcio! Happy to see you're finding it useful ^^
This is the part I keep running into — the useful data in a reel is almost never in the caption, it's spoken or on screen. Getting it back as JSON with a schema I describe in plain language is what makes this click for me as a dev. Curious how you handle it when the audio and the on-screen text disagree — which source wins? Congrats on the launch, Unai.
@jcbuilds thanks! As for the contradiction case, an AI model will reason through and will decide what the best answer is on a case-by-case basis ^^