Every AI video tool gives you a model menu. Almost none give you a director.
Nolan is built around a simple idea: you describe what you want — a product shot, a talking ad, a six-shot teaser — and an AI agent plans it, picks the right frontier model and writes the prompt. You see what it will cost, approve it, and it generates. No node graphs, no model menus to study.
Updated October 2026: Nolan is now chat-first. The Studio storyboard, the built-in editor and Enhanced mode described in the original version of this post have been removed.
What's inside
- Chat, Agent mode — describe a shot or an image and refine it by talking. The agent plans it and proposes each generation on an approval card with its estimated cost; nothing runs until you approve.
- Chat, Manual mode — full control: choose the model, duration, resolution and aspect ratio, and your prompt goes straight to the model.
- Characters — create a character once, approve its look, and keep the same face across everything you generate. Save it to your Library to use it in any project.
- Projects and Library — each project keeps its brief, script, characters and results together; what you save to the Library is ready to reuse anywhere.
The models
Nolan runs a roster of frontier models behind one interface: Seedance 2 for native dialogue and lip-sync, Kling 3.0 for stylised multi-shot sequences up to 4K, Wan 2.7 for text/image/reference-to-video and instruction-based video editing, plus GPT Image 2 and Nano Banana Pro for stills. The agent picks the best one for each job — or you choose, in Manual mode.
Simple pricing
1 credit = $0.01, and you pay only for what you generate: you see an estimate before every generation, and a failed generation is never charged. Credits come in one-time packages. Try it now.