Turn a topic into a documentary nobody has to film.
Write one line about what you want covered. FrameVideo researches it, writes the script, narrates it, finds or generates every visual, and cuts the whole thing together.
Long-form and vertical · 11 visual styles · your own cloned voice
Three decisions, then it builds itself
- 01
Give it a topic
One line is enough. Add your own angle and the script treats it as authoritative — your take wins over the model’s, and over the research it gathers.
- 02
Pick a look
A niche sets the writing voice and pacing. A visual style sets the imagery. Horizontal for long-form, vertical for shorts.
- 03
Get a finished video
Script, narration, sourced or generated visuals, word-synced captions, music and sound design, cut together and rendered. Download the MP4.
11 visual styles
Every frame below was produced by the same pipeline your videos use — real output, not mockups.
Built for a kind of channel, not “video” in general
Each niche changes how the script is written, which scenes get used, and how the video is paced — a told story is not a documentary with different footage.
History / Documentary
Measured, authoritative narration. History and human stakes come first.
Story / Folklore / Legend
A tale told scene by scene, not explained — characters act, and every visual shows the moment it reaches.
True Crime / Mystery / Storytelling
Tense, investigative storytelling — reveal facts like clues, build toward a question.
Finance & Business
Credible, data-driven, calm authority — like a business news segment.
Science & Space
Curious, wonder-driven, precise — explains the mechanism before the implication.
Health
Clear, trustworthy, human — never alarmist.
Top 10 / Listicle
A punchy, ranked countdown — high energy, rank-to-rank momentum.
The details that make it watchable
Images land on the word
Each visual is anchored to the exact word in the narration it illustrates, so the flag appears as the narrator says the country — not wherever an even split happened to put it.
Every image is checked
A vision model looks at each candidate before it is used: is it on topic, where is the subject, and would cropping destroy it. Rejected images are not quietly reused.
Characters stay themselves
Story videos build a cast list with fixed appearance descriptions, folded into every generated image that names them, so your protagonist is recognisably one person across scenes.
Your voice, optionally
Use a stock narrator or clone your own from a short sample, so a channel keeps one voice across everything it publishes.
Scored, not just narrated
Mood-matched music that ducks under the voice and swells in the gaps, plus sound design on cuts and reveals — mixed at measured levels, not guessed.
A channel look you keep
Save a kit — colours, fonts, transitions, voice — and every future video inherits it, so a channel looks like one channel.
Questions
- What exactly do I get?
- A finished MP4 with narration, visuals, captions, music and sound design. Horizontal for YouTube, vertical for Shorts, Reels and TikTok. You download the file and publish it yourself.
- How long can a video be?
- Up to about ten minutes. Longer videos are structured into acts with their own mini-hooks rather than one flat block of narration.
- Where do the visuals come from?
- Depends on the style you pick. Real Footage searches licensed and public-domain archives and credits every source. Every other style generates its imagery, which is what lets a story depict moments no camera was present for.
- Can I use the videos commercially?
- Yes. Sourced images carry their attribution into the video’s credits and metadata so the licensing travels with the file.
- Do I need to edit anything afterwards?
- No. The output is publishable as-is. You get the MP4, not a project file to finish.
Start with one sentence
Pick a topic you would watch a video about. See what comes back.
Start a video