TwelveLabs is launching today Marengo 3.0, the most powerful embedding model for video understanding.
I had the opportunity to collaborate with their team on this launch and wanted to share some behind-the-scenes on how we prepped it.
Setting up the launchpad
We started by working on the required inputs to submit a new product on Product Hunt:
- Product name: 40 characters max
- Product tagline: 60 characters max
- Product description: 500 characters max
- Thumbnail: recommended size = 240*240px; filetype = JPG, PNG, or GIF; maximum size = 2MB
- Image gallery: recommended: 3 or more images; optimized size = 1200*630px
- Launch tags: up to 3 tags
Pro tip: You can find the required inputs on GitHub.1
We had a clear messaging, so we focused effort on the image gallery.
Showing the product
On Product Hunt, the image gallery tells a story, visually. It starts with an intro, i.e. the Open Graph image. IMHO it should finish with an end, i.e. your call-to-action.
We brainstormed:
How to tell a compelling story in 3 or more images?
Below are the requirements we defined:
- Image 1: Open Graph image
- Images 2-6: Feature highlights
- Image 7: Call-to-action
We used a 1200:630 ratio for optimal display.
Above all, because TwelveLabs targets a technical audience, we wanted to show the product. We framed product screenshots and added a headline and subheadline for context.
Show, don't tell.
To recap: We had a clear messaging. We worked on a compilling visual story. Next: Scheduling the launch.
The best launch day != The right launch day
We thought:
What's the best day to launch [on Product Hunt]?
In their launch guide,2 Product Hunt puts it simply: The best day to launch is when you're ready.
Today is AWS re:Invent 2025, the ultimate week for builders, dreamers, and disruptors, and TwelveLabs has geared up for this event with a series of announcements.3
In this context, we decided to launch today, Monday, December 1st. First day of the week, first day of the conference. Hopefully this decision will give the launch maximum exposure.
Wrapping up
That's it! TwelveLabs is launching today Marengo 3.0, the most powerful embedding model for video understanding.
Top comments (0)