Timeline questions usually get a vague answer industry-wide. Here are real numbers from real projects, and what actually determines whether an AI film production takes days or weeks.
Why timeline questions get vague answers
“It depends” is technically true and practically useless. What it depends on is knowable in advance: whether the project is a full film or a sequence integrated into existing footage, how many scenes it needs, and how much historical or brand accuracy it has to hold to.
Pre-production time
Research, references and character or world design happen before any generation starts. A project with real historical accuracy requirements, like St. Nicholas, needs more pre-production time than an original concept piece where the world’s rules are being invented rather than researched.
Generation and consistency passes
Each scene is a generation-and-review cycle, checked against the locked character and world reference before it’s accepted. More scenes and more consistency requirements both add real time here — it’s the single biggest variable in any timeline estimate.
Edit and delivery
Once scenes exist, edit, color, sound and final delivery follow — proportional to runtime and how much finishing work the project calls for, whether that’s AI narration, a specific grade, or platform-specific cuts.
Real timeline references
St. Nicholas — a 1:32 biography across four scenes with AI narration — took 27 days. Zarafel — three integrated sequences of 20 seconds or more, slotting into an existing live-action film rather than building a world from nothing — took 7. A full film and an integrated sequence are genuinely different amounts of work, and we quote them that way. See our full pricing breakdown for how this maps to cost, or read about our AI film production process.
Leave a comment