One render is capped at 180 seconds. A longer film is written in chapters, rendered as segments, and stitched together afterwards — and the stitching reuses clips that already exist, so it is free and takes minutes.
A ten-minute job would occupy the renderer for around half an hour, and a failure at minute 28 loses everything. Segments fail small and retry cheaply. They are also better to watch: a 600-second job gets 40 scenes of 15 seconds each, and fifteen seconds of slow zoom on one picture is unwatchable.
Segments of about 150 seconds keep scenes near the nine seconds the pipeline is tuned for.
A film is written scene by scene, with questions asked before a word is committed. That is a conversation, not a web form with a dropdown. So the app connects to Claude as an MCP connector — one URL, added once, no key to paste:
https://automatedvideoapp.com/mcpAdd it under Settings → Connectors → Add custom connector, sign in when prompted, and Claude gains ten tools: fetch the production guide, see what each option costs, create a video, check its progress, list its scenes, drop a scene, set the end card, re-cut, list your uploaded images, and join finished videos into one longer film.
From then on you ask in plain sentences. Claude reads the production method first, asks what it needs, writes the script to the right length, and renders it — then edits it with you afterwards, for free, because re-cutting reuses clips that already exist.
That is the difference from a trailer generator you fill in: the thing that plans the film is the same thing you are talking to, and it can go back and change one scene without rebuilding the rest. More on automating video with Claude →
These are different things and it helps to keep them separate:
Every segment must begin and end on a clean beat. It is rendered separately, so a mid-sentence boundary is audible at the join.
Aspect, voice, look and image source must match across all segments, or the join will look like different films spliced together. Carry any character description and the colour palette verbatim from the first segment to the last — otherwise chapter three will not look like chapter one.
Once every segment has finished, they are stitched into a single film in play order. This re-uses the clips already rendered: no generation, no Pollen, and a couple of minutes rather than another hour. The original segments stay untouched and still playable, so replacing one and re-joining is cheap.
The limits are twelve segments and 240 scenes per join — about 36 minutes. Beyond that, join in batches and then join the results.
Roughly one hour of rendering per twenty minutes of film, and the renderer handles one job at a time. Worth saying out loud before starting rather than discovering at minute forty.
Watch each segment as it lands. A bad one caught at minute three is cheap; the same problem found at the end is not.
Ten minutes of a shifting protagonist is unwatchable. Ten minutes of a consistent presenter is a normal format. An avatar presenter is the only real fix for face consistency at length — it needs your own HeyGen key, which the app does not supply, so connect it before writing a script rather than after.
Length is not the same as a screenplay. No current tool holds narrative coherence across ninety minutes, and it is kinder to know that at the start than to find it out at the end.
Connect Claude and plan a long filmNext: make it look cinematic, or start with a trailer, or publish a series on a schedule.