Higgsfield open-sources every prompt and the full production playbook behind its 95-minute AI film, Hell Grind
- They didn't just release the film. They released 115,000+ generation records. Every single prompt is viewable and copyable.
- The most valuable part: the 12-line "technical foundation" shared by all shots — from skin pores to 60:30:10 color — is copy-paste ready.
- That mountain of prohibitions is a user manual read backwards. The model adds extras, duplicates people, and improvises dialogue.
They laid the entire production ledger open.
Higgsfield has open-sourced every prompt and production method behind its 95-minute AI feature film, Hell Grind. All prompts and assets are now public. The film was made for $500,000, screened at the Cannes Market, and has been covered by the Wall Street Journal, Variety, and BBC News.
Here's what's public: 115,446 generation records organized into 108 folders by scene. Clicking any record reveals the full prompt used at the time, the model, the parameters, and a direct link to download the asset. The full film is embedded at the top of the page — 95 minutes 6 seconds, free to watch.
This article isn't a review. It does one thing: reverse-engineers the production method from 40,000+ prompts and hands you the parts you can copy, verbatim.
Every shot used the same 12-line "technical foundation".
When we split those 40,000+ prompts into individual lines and counted repeats, the first thing that surfaced was this. These 12 lines sit at the end of nearly every prompt. The most frequent one appears 8,015 times.
It functions like a shooting spec for the entire crew: regardless of the shot, the standards for image quality, lens, light, color, skin, physics, performance, composition, continuity, frame rate, and audio are locked in.
Style: 8K IMAX. Photorealistic — no 3D render, no game engine, no game-cutscene aesthetic. Cinematography: Emmanuel Lubezki × Roger Deakins. Camera: Physical cine lens. 180° shutter motion blur. Lighting: Natural light only — contre-jour backlight, camera on shadow side, atmospheric haze throughout. Key light from sky and windows only. Color: 60:30:10 — dominant / secondary / accent. Skin: Pore-level realism — vellus hair, asymmetric moles, capillary flush, pore-shadow matching on-set light. Physics: Gravity and inertia respected — mass has real weight, correct contact shadows. No floating props. Acting: Hollywood — micro-pauses before reactions, precise eye-line, wet living eyes with catch-lights, visible breath and chest rise. Composition: Rule of thirds + golden ratio. Every person moving from frame one. Continuity: Characters, props, environment identical across every cut. No identity drift. Technical: 24fps smooth motion. 8K detail. No jitter. Audio: Environmental SFX only. No music. No subtitles.
Here's what each line does and how often it appears in the archive:
| Line | What it prevents | Count |
|---|---|---|
| Style: 8K IMAX, photorealistic, no 3D render, no game engine, no game-cutscene aesthetic | The most typical "plastic" look in AI video comes from looking like game CG. Naming and banning these three is more effective than writing "be realistic" a hundred times. | 5,820 |
| Cinematography: Emmanuel Lubezki × Roger Deakins | Using cinematographers' names as style anchors. More precise than adjectives. | 5,219 |
| Camera: Physical cine lens, 180° shutter motion blur | The 180° shutter rule is the physical basis of the film look. Locking it in prevents that overly sharp, electronic look where every frame is crisp. | 6,693 |
| Lighting: Natural light only, contre-jour, camera on shadow side, atmospheric haze throughout, key light from sky & windows only | This is the most critical line. AI's default is flat, bright lighting. This hands control back to the weather and windows. | 4,942 |
| Color: 60:30:10 dominant / secondary / accent | A classic design ratio that prevents visual chaos. | 6,444 |
| Skin: Pore-level realism — vellus hair, asymmetric moles, capillary flush, pore-shadow matching on-set light | Names the four things skin needs. "Realistic skin" is a platitude; these are executable instructions. | 6,822 |
| Physics: Gravity & inertia respected — mass has real weight, correct contact shadows. No floating props. | Objects floating and feet not touching the ground are AI video's most common failure points. | 6,490 |
| Acting: Hollywood — micro-pauses before reactions, precise eye-line, wet living eyes with catch-lights, visible breath & chest rise | Deconstructs "acting human" into four executable details. | 4,146 |
| Composition: Rule of thirds + golden ratio. Every person moving from frame one. | The second half is key: without it, AI tends to give you a group of people standing still. | 7,252 |
| Continuity: Characters, props, environment identical across every cut. No identity drift. | The lifeblood of a feature film. | 7,706 |
| Technical: 24fps smooth motion, 8K detail, no jitter | Frame rate and quality baseline. | 8,015 |
| Audio: Environmental SFX only, no music, no subtitles | The model will score your film and add subtitles on its own. You have to turn it off explicitly. | 7,513 |
A 15-second shot's prompt has seven sections before the foundation.
The technical foundation is shared. The large block before it is shot-specific. The median prompt length is 16,501 characters — roughly 3,000 English words. The longest is 39,801 characters. All of that text follows a fixed skeleton.
Current injuries, where clothes are torn, facial expressionLargest
What state this shot picks up from34%
Director's intent, an anchor for the model16%
Who is where, distance, direction37%
Even silent shots need wind and footsteps detailed36%
15 seconds split into 6 beats, one sentence each48%
What must be there, what must never appear33%
Stats based on 4,748 prompts over 3,000 characters; sections can co-exist so percentages don't sum to 100%
Why the first section is longest: the model has no memory
Video models don't know what happened in the previous shot. They don't know this person was stabbed three minutes ago. If you don't write it, the model makes something up. So every shot re-establishes the character's current state from scratch — to this level of detail: