--- title: README emoji: 📼 colorFrom: gray colorTo: red sdk: static pinned: false ---

Darkframe AI publishes horror-native weights and tooling: we are starting with found footage and then branching off into all the subgenres we explore like body horror, zombies, vampires, folk horror, and many more.

darkframe.ai · Models & collections

### Hub releases | | | | --- | --- | | **DARKFRAME — Lost Footage v1** *(preview)* | [Model card](https://huggingface.co/darkframeai/DARKFRAME-LostFootage-v1) · Early-access checkpoint for night-vision cast, VHS glitches, grain, and claustrophobic “discovered recording” stills. Apache-2.0. |

Preview weights may change—see the model card for the full disclaimer. Prompting notes, gallery, and file layout live on each model page.

### Visual language Representative stills (training / eval references for Lost Footage v1):
Found footage corridor, institutional horror, tape degradation CRT glow, abandoned interior, analog noise Found footage mood board, grain and dread
### For filmmakers & prompt engineers We optimize for **genre-native** output: handheld framing, wrong lighting on purpose, chromatic bleed, and locations that feel like **evidence**—not hero shots.

New checkpoints and datasets land here first. Watch this org for releases.