Lineart Anime LoRA for FLUX.2 Klein 9B

• What it does: Generates clean, professional anime-style lineart illustrations from text prompts
• Base model: FLUX.2 Klein 9B (black-forest-labs/FLUX.2-klein-9B)
• Trigger words: Lineart, LineAniAF (HF) / AnimedREDAF, Anime (Civitai)
• Format: SafeTensor LoRA, 81 MB, Apache 2.0 license

Overview

This LoRA adapter was trained by ArtificialGuyBR to add anime lineart capabilities to FLUX.2 Klein 9B. The training was made possible by GPU compute sponsored by Redmond.AI. It excels at producing consistent black-and-white lineart with clean lines, proper shading, and anime/manga aesthetics — suitable for character art, scene illustrations, and stylized designs.

Trigger Words & Prompting

Use the trigger words at the start of your prompt:

• HuggingFace: Lineart, LineAniAF
• Civitai: AnimedREDAF, Anime

Example prompts from the model card:

• Lineart, LineAniAF girl in traditional kimono, cherry blossom garden, petals floating, serene expression, Japanese style, detailed patterns
• Lineart, LineAniAF schoolgirl with long hair, classroom background, books on desk, sunlight through window, nostalgic vibe, clean lines
• Lineart, LineAniAF young woman playing acoustic guitar, park bench, autumn leaves falling, peaceful moment, soft shading, manga style

How to Use

Recommended: ComfyUI

1. Download the SafeTensor file (81 MB)
2. Place in your ComfyUI models/loras/ folder
3. Add a LoRA loader node to your workflow
4. Set strength around 0.7–1.0 (start at 0.8)
5. Include trigger words in your positive prompt

The model also has variants for ZImageTurbo and Qwen Image base models on Civitai — download the version matching your base model.

Example Outputs

The images demonstrate consistent lineart quality: clean black lines on white, proper anime proportions, effective cross-hatching and shading, and recognizable manga/anime aesthetics.

Technical Details

Property | Value

|---|---

Base model | black-forest-labs/FLUX.2-klein-9B Type | LoRA adapter Format | SafeTensor Size | 81.15 MB License | Apache 2.0 Trigger words | Lineart, LineAniAF Trainer | ArtificialGuyBR GPU sponsor | Redmond.AI Civitai downloads | 422 Civitai likes | 18 Verified | May 2026

Common Errors and Fixes

Error | Cause | Fix

|---|---|---

No style change | Trigger words missing | Add Lineart, LineAniAF to prompt start Overcooked lines | LoRA strength too high | Reduce to 0.5–0.7 Color artifacts | Base model expects color | This LoRA is trained for B&W lineart; expect monochrome output Wrong base model | Using non-Klein FLUX | Match LoRA variant to your base model (Klein 9B, ZImageTurbo, or Qwen)

FAQ

Q: Does this work with FLUX.1 Dev/Schnell? A: No, this LoRA is trained specifically for FLUX.2 Klein 9B. Using it with other FLUX versions may produce poor results.

Q: Can I use it commercially? A: Yes, Apache 2.0 license permits commercial use with attribution.

Q: What LoRA weight should I use? A: Start at 0.8. Range 0.6–1.0 typically works well.

Q: Are there other base model versions? A: Yes, Civitai hosts variants for ZImageTurbo and Qwen Image. Download the matching variant.

Sources

• HuggingFace model card: https://huggingface.co/artificialguybr/LineartAnime-Redmond-Flux9BKlein
• Civitai model page: https://civitai.com/models/2422596
• Redmond.AI compute sponsor: https://redmond.ai/
• ArtificialGuyBR portfolio: https://artificialguy.com/