Behind the scenes

How an AI film really comes to life.

No push of a button, no accident. Behind every image is a craft of six stages – and for each stage the best tool. Here we show you how we work and with which tools.

Virtuelle Produktion – Charaktere und Szenen
The workflow

Six stages from idea to image.

Each stage has a clear goal – and a sign-off point with you. So out of hundreds of generated options, what remains is exactly the one film that was meant.

1

Idea & screenplay

Concept, story beats, dialogue. The story stands before a single image exists.

Text & story models
2

Storyboard & alignment

We visualise every key scene as an image and align it with you – before anything is animated.

Image models for frames
3

Image design & look

Lighting, colour, era, visual language. We define the film's visual fingerprint.

Cinematic image models
4

Creating characters

Characters that look the same in every shot – via references, style anchors and fine-tuning.

Consistency & LoRA
5

Animating images

Frames become shots: camera moves, motion, dialogue, timing – consistent across sequences.

Video models
6

Music, sound & edit

Composed score, sound design, voice, color grading and final cut. This is where it becomes cinema.

Audio & post
Our principle

For every problem the right tool.

We are tied to no platform. The landscape of AI tools changes monthly – so we choose the best available model per shot and combine their strengths. What counts is the image on the screen, not the logo behind it. This independence is our greatest craft advantage.

Our toolbox

The tools we build with.

A selection of the models we use in 2026 – and what each is best for. We update this stack continuously as better tools appear.

Tool
What we use it for
Strength
ChatGPT Text & story
Ideas, screenplay, dialogue, research and structure.
Narrative & logic
Google Imagen 4 Image
Photorealistic concept images, reliable text in image.
Realism & text
Nano Banana Pro / 2 Gemini 3 · Image
Fast, consistent character and scene images, variations.
Character consistency
Midjourney V7 Image
Artistic, cinematic looks and moodboards.
Aesthetics & style
Recraft V4 / LoRA Consistency
Recurring characters, brand looks and style anchors.
Brand & style fidelity
Grok Imagine Video
Fast previz and affordable shots with native audio.
Speed & cost
Kling 3.0 Omni Video
Multi-shot sequences with dialogue and frame-accurate control.
Control & sequences
Seedance 2.0 Video
The most cinematic shots, with reference and multimodal support.
State of the art
Google Veo 3.1 Video
Strongest all-around quality with synchronized sound.
All-around & audio
Score & sound design Audio
Composed music, foley, voice and final mix.
Emotion in the ear

As of 2026. Tool names belong to their respective providers; the selection per project depends on task, budget and look.

Why this matters

The craft lies in the choice.

A good model doesn't make a good film – just as a good camera doesn't make a good director. The art lies in knowing which tool best solves which scene, how to keep characters consistent across dozens of shots, and how many single images become a film with rhythm and breath.

That's exactly what we're here for. You don't get "AI output", but a finished film – with a team that owns the whole path from the first idea to the final master.

FAQ

Frequently asked questions.

No. We are model-agnostic and choose the best available tool per shot – the selection is our craft advantage.

Through references, style anchors and fine-tuning we keep characters consistent – shot by shot.

A finished film. We own the whole path from the idea to the final master.

Curious how your film would come to life?

We'll show you the path on your concrete project – from idea to cut.

Discuss a project