Explainer videos, through the API
Every product needs the same video. Thirty to ninety seconds: here's the problem, here's what we do about it, here's how it works, here's the proof, here's what to do next. It goes on the homepage, opens the sales deck, launches the feature. And it is, reliably, the thing that doesn't get made — because it means a brief, an agency, three rounds of notes and six weeks.
As of today, Second Studio makes it from an API call.
Three inputs, one video
A brand kit. You save it once: what the company does, the tone you write in, five colors, a font,
your logo, and the product screenshots we're allowed to show — each with a one-line description so
the writer knows what it's looking at. That's POST /v1/brands, or create_brand over MCP.
A script. Give us a brief — audience, pain, the two or three things that matter, any real numbers, the call to action — and the studio writes the scenes. Not prose: a list of typed scenes (title, steps, bullets, stat, quote, image, compare, call to action), each with a headline, its on-screen content, and the narration a voice would read. It comes back as JSON. You change a word, reorder two scenes, cut the one you don't like, and send it on. Or skip us and write the scenes yourself.
A render. POST /v1/explainers turns scenes plus brand into an MP4: 16:9, 9:16 or 1:1 from the same
script, with captions, narration where it's enabled, a music bed if you give one.
What makes it dependable
The model writes words. It never writes layout, picks colors or draws an interface. Each scene type has one fixed, brand-aware component, so the same script and brand produce the same video every time; re-rendering after a one-line edit changes one line. Screenshots in a video come only from the images in your brand kit — we don't generate UI, ever. And the writer is held to your brief: no invented statistics, customers or awards. If you don't give it a number, it won't make one up.
The script step is 2 credits, a render is 12. A typical explainer, from brief to file, is under five minutes.
Where it fits
Second Studio already lets an agent analyze a reference video's editing style and rebuild it from your own clips, record real product UI, generate characters, images and shots, and post to connected social accounts. Explainers fill the gap in the middle: the structured, branded, say-what-it-does video that most teams need first and make last.
Read the explainer docs, or try the flow from any MCP client:
create_brand → write_script → create_explainer.
If you'd like access, ask. Keys are approved by hand, and we read every request.