🔪NSO Pixel Style - Needy Streamer Overload Art Style LoRA👾
A style LoRA trained to replicate the distinctive dithered pixel-art aesthetic of Needy Streamer Overload (NEEDY GIRL OVERDOSE). Trained on Anima.
⚡Trigger:
nso style, pixel art, dithered, jagged edges, retro 90s aestheticUse the trigger phrase together for best results - it's pinned as a fixed block during training, so keeping it together (rather than splitting the words apart) gives the most consistent output.
What this does well:
Reproduces the game's dithered/halftone shading and hard pixel-grid linework
Works across different characters and poses, not just the source game's cast
Holds up cleanly at ~1024×1024 generation
⚙️Recommended settings:
Strength: 1.0
Resolution: 1024×1024
Steps: ~25, CFG ~4.0 (Anima defaults)
Heads up: this LoRA was trained on a dataset that includes some intense/dark imagery consistent with NSO's actual narrative content (the game deals with streamer burnout and darker themes). Style transfer means it can replicate that range if prompted toward it - keep that in mind with your own generations.
This is a fan-made project, not affiliated with or endorsed by the developers of Needy Streamer Overload. Please don't use this for commercial work.
Description
v1.1 - retrained
The previous version was looking crooked or broken on other Anima checkpoints. v11 is retrained from scratch on plain Anima Base v1.0. It now gives the same look on all Anima checkpoints from official distribution and probably more.
Trigger / prompt start:
nso style, pixel art, dithered, retro 90s aesthetic
Recommended strength: 1.0 (lower it for a softer look)
Works with: Anima Base, Anima Turbo, Anima Aesthetic and their merges
Changes vs previous version:
- Consistent style across all Anima models
- Cleaner lines and better anatomy, faces and hands
- Works for Ame-chan, KAngel and original characters
- Slightly less heavy dithering than the old version; add "dithered" or raise the strength if you want it chunkier
Training: 245 images (in-game frames from Needy Streamer Overload + curated NSO-style images), WD14 tags, rank 32, 768px, 8 epochs, OneTrainer.