Skip to content
WorkflowVerifiedExternal
Download

Broken Anima — v1.0

Basically, it generates in low resolution first and then finishes at the original resolution. It uses just the basic latent upscale, so it's a bit unstable—but hey, it's just for my own use anyway.

Sourcecivitai.com/models/2664123
TaskUpscalingFamilyAnimaVersionRelease 2991565 — ANIMA.jsonEstimated VRAMEstimate unavailableBgrade

Runtime profile

Access
External workflow
Status
Verified
Version
ANIMA.json
Source
Civitai
Updated
Oct 1, 2026

Source excerpt

No release description captured.

Requirements

Models

5

  • anima-turbo-lora-v0.2.safetensors

    File specified

    LoRA · 142 MB · Civitai · 2560840

  • anima-turbo-v1.0-gpt.safetensors

    Not resolved

    Checkpoint · Unknown

  • gpt-image-2_anima-base1_v1.safetensors

    Not resolved

    LoRA · Unknown

  • qwen_3_06b_base.safetensors

    Not resolved

    Text encoder · Unknown

  • qwen_image_vae.safetensors

    Found

    VAE · 242 MB · SAFETENSORS · BF16 (tensor header) · Civitai · 2641326

AI-generated commentary

AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.

Purpose

  • Broken Anima is a workflow for generating images from text conditioning and an empty latent image; it starts at low resolution, applies basic latent upscaling, finishes at the original resolution, and produces a decoded image plus a preview.

Flow

  • Named processing components include KSamplerSelect, SamplerCustom, SplitSigmas, LatentUpscaleBy, VAEDecode, and PreviewImage.

Inputs and outputs

  • The named inputs are text conditioning and an empty latent image; the named image outputs are a decoded image and a preview.

Setup

  • This is version v1.0, published by M8rr, and based on Anima.

  • No license is listed.

Dependencies

  • The workflow lists these model files: anima-turbo-lora-v0.2.safetensors; anima-turbo-v1.0-gpt.safetensors; gpt-image-2_anima-base1_v1.safetensors; qwen_3_06b_base.safetensors; and qwen_image_vae.safetensors.

  • Named nodes include CLIPLoader, CLIPNegPip, CLIPTextEncode, ConditioningZeroOut, EmptyLatentImage, KSamplerSelect, LatentUpscaleBy, and LoraLoaderModelOnly.

  • Other named nodes are LTXVScheduler, PreviewImage, PrimitiveStringMultiline, SamplerCustom, SplitSigmas, StringConcatenate, UNETLoader, VAEDecode, and VAELoader.

Configuration

  • To inspect sampling and upscaling controls, start with KSamplerSelect, SamplerCustom, SplitSigmas, and LatentUpscaleBy.

    Suggestion · not verified

Cautions

  • The source describes the basic latent upscaling as “a bit unstable”; review the output with your own inputs.

    Suggestion · not verified

  • Before using the workflow, verify that its listed model files and nodes are available in your setup.

    Suggestion · not verified

Does this need to be edited?

Sign in to send an edit request.

Sources

1 source
CivitaiSourceVersion 2991565
civitai.com/models/2664123

Verified Oct 1

Model artifacts

1 artifact

brokenAnima_v10.zip

zip · 7.14 KB · SHA-256 a6858e15ff74…cb90 · Civitai

Download file

Source excerpts

3 excerpts
NoteOct 1civitai.com/models/2664123

Source context: 168 downloads · Type Workflows · Base model Anima

NoteJul 17civitai.com/models/2664123

Source context: 113 downloads · Type Workflows · Base model Anima

NoteJul 11civitai.com/models/2664123

Source context: 112 downloads · Type Workflows · Base model Anima