UltraRealistic LoRA Project (ULP) v2: candid, amateur-style photos on Flux
Fast Facts:
The UltraRealistic LoRA Project (ULP) for FLUX.1 Dev by Danrisi aims at more 'live' images: dynamic poses, real emotions and a slightly amateurish photographic quality that reads as authentic. It can still produce polished images, but its gift is the candid snapshot.
Trigger phrases from the author include amateurish photo, low lighting, in motion, overexposed, underexposed, GoPro lens, eerie atmosphere, smeared background and smeared foreground; each nudges a specific photographic flaw.
Reference
Name | Type | What it is | |
|---|---|---|---|
Trigger phrases | text | amateurish photo, low lighting, in motion, overexposed, underexposed, GoPro lens, smeared background | |
LoRA weight | 0.6-1.0 | 0.8 is a good default. | |
Guidance | 2-3 | Low guidance keeps skin and light natural. | |
Steps | 25-30 | Standard Dev. | |
| Base | FLUX.1 Dev | Also works on Krea Dev. |
Step by step
- Start with amateurish photo, then describe the moment as a snapshot.
- Add one flaw phrase that fits (low lighting for a bar, in motion for a street).
- Set weight 0.8, guidance 2.5, 28 steps.
- Avoid words like masterpiece and 8k.
- Pick frames with imperfect framing; that is the point.
Examples
Snapshot
amateurish photo, a grainy snapshot of a woman in a worn red shirt sitting on an old basement floor looking up, low lighting, slight motion blur
ULP 0.8, guidance 2.5
Party flash
amateurish photo, direct flash, three friends in their twenties squeezed into a photo booth, overexposed faces, dark background
Creative idea: fake found footage
amateurish photo, GoPro lens, a cyclist's point of view on a foggy forest trail at dawn, smeared foreground, underexposed
Tips
- Describe what the person is doing, not how they pose; ULP responds with natural gestures.
- Flash photography prompts (direct flash, red eye) come out convincingly.
- Combine with SRPO at 0.6 for better skin while keeping the candid feel.
- Anatomy can wobble in extreme poses; the author notes it. Lower the weight or regenerate.
Troubleshooting
Faces look too perfect
Why it happens
Guidance too high or weight too low.
How to fix it
Guidance 2, weight 0.9, add amateurish photo.
Hands or limbs break
Why it happens
Dynamic poses push anatomy.
How to fix it
Lower the weight to 0.6 and describe a simpler pose.
Questions
Does it work with Flux Krea?
Yes; Krea's natural grade and ULP's candidness complement each other.
Is a trigger word mandatory?
No, but the phrases steer the look reliably.
Links and sources
Models in this guide
Written by
Runs this ship. Has opinions about VAE decode times that nobody asked for. Believes weights want to be free and that any model you cannot download is a subscription with extra steps. Will not be taking questions about the 400 GB of checkpoints on the NAS.
All guides by Captain
Related guides
Shou Xin pencil sketch LoRA: minimalist graphite drawings on Flux
Oct 3, 2026
Flux SRPO LoRA: better skin and portraits for FLUX.1 Dev
Sep 30, 2026
Shuttle 3.1 Aesthetic: fast, pretty Flux Schnell images in 4 steps
Sep 29, 2026
Flux Ghibsky Illustration: serene painted skies on Flux
Oct 3, 2026
FaeTastic Details for Flux: fantasy detail and shimmer
Sep 29, 2026
More guides
What is a LoRA and how to use one in ComfyUI, Forge, Automatic1111 and on the cloud
May 20, 2026
Flux vs SDXL vs SD 1.5 vs Qwen vs Z-Image: which model family to use in 2026
May 22, 2026

Flux vs SDXL in 2026: Which Should You Use (Including Flux.2)?
Oct 16, 2026
Flex.1 Alpha by Ostris: the open, fine-tunable Flux cousin
Sep 30, 2026
How to train a LoRA at home (and when to rent a GPU instead)
Aug 5, 2026
