> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dopaminegirl.com/llms.txt
> Use this file to discover all available pages before exploring further.

# How to Upload LoRA

> Bring Flux, Stable Diffusion, and Wan LoRAs from CivitAI into DopamineGirl without guesswork.

Want to add your favorite LoRA to DopamineGirl? This guide walks through the exact filters to use on CivitAI, how to double-check the files, and the two upload paths (direct upload vs. URL fetch). We’ll demo the flow with Flux1 Dev, then call out what changes for Wan v2.2 video LoRAs and Stable Diffusion v1.5.

## Before You Start

* Check the [public LoRA library](https://dopaminegirl.com/lora) first—your model might already be live.
* Only upload LoRAs that provide a `.safetensors` download and are not marked **Early Access**.
* Direct uploads support files up to **400 MB**; larger models should use the URL option.

## Step 1. Find the LoRA on CivitAI

Head to the [CivitAI models page](https://civitai.com/models) and open the filter drawer. Set the filters based on the base model you need:

### Flux1 Dev LoRA

1. Under **Model Types**, choose `LoRA`.
2. Under **Base Model**, pick `Flux .1 D`.

![](https://asset.dopaminegirl.com/docs/0N9DXMFT2E98M.webp)

That narrows the list to community-rated Flux LoRAs:

![](https://asset.dopaminegirl.com/docs/0N9DY9J9K78PC.webp)

Click a result to inspect its details:

![](https://asset.dopaminegirl.com/docs/0N9DYW0TK734H.webp)

Confirm there is at least one `.safetensors` file listed in the downloads section and that the version is public. Copy the model URL—you’ll need it if you want DopamineGirl to fetch the file for you.

### Z Image

Filter for `LoRA` + `Z Image Turbo`:

![](https://asset.dopaminegirl.com/docs/0P2AR685MSWFN.webp)

### Wan v2.2 LoRA (Video)

Use the same `LoRA` model type filter but change the base model to `Wan v2.2`:

![](https://asset.dopaminegirl.com/docs/0N9E1EF5JE77Q.webp)

Wan video LoRAs ship in two flavors:

* `t2v` (text-to-video)
* `i2v` (image-to-video)

Upload the variant that matches how you plan to generate. Each variant also comes in high-noise and low-noise versions—choose the **high-noise** file because that’s what DopamineGirl supports today.

![](https://asset.dopaminegirl.com/docs/0N9E217F3Z9VK.webp)

### Stable Diffusion v1.5 LoRA

Filter for `LoRA` + `Stable Diffusion 1.5`:

![](https://asset.dopaminegirl.com/docs/0N9E2DMPVZ9TC.webp)

Pick the style you want and open the model page to review the downloads just like you did for Flux.

## Step 2. Upload to DopamineGirl

Once you have the LoRA picked out, switch to the DopamineGirl upload form. You have two paths:

1. **Direct Upload** – Download the `.safetensors` file from CivitAI and drop it into the uploader.\
   ![](https://asset.dopaminegirl.com/docs/0N9DZJF9QZ9DC.webp)\
   Use this when the file is under 400 MB.

2. **CivitAI URL Fetch** – Paste the model URL into the input and let DopamineGirl pull the file for you.\
   ![](https://asset.dopaminegirl.com/docs/0N9E03D92EAGS.webp)\
   Example: [`https://civitai.com/models/915918/...`](https://civitai.com/models/915918/moriimee-gothic-niji-or-lora-style?modelVersionId=1189748)

After the upload finishes, the LoRA shows up in your collection—and if it’s public, the community can use it too.