LoRAs API¶

Endpoints for managing the style LoRA catalog. Style LoRAs are kept in a separate catalog (catalogs/loras.json) from the models catalog. Download and delete operations share the same endpoints as models (use the model filename with /api/admin/models/download/{filename} and DELETE /api/admin/models/{filename}).

GET /api/admin/loras¶

List all style LoRAs from the catalog with download/presence status.

Auth: Required

Response: 200 OK

{
  "version": 2,
  "date": "2026-03-20 10:00",
  "stats": {
    "queued_count": 0,
    "downloading_count": 0,
    "present_count": 5,
    "total_count": 15,
    "global_speed": 0,
    "models_bytes": 2147483648,
    "free_bytes": 42949672960
  },
  "categories": [
    {
      "id": "wan-loras",
      "name": "WAN LoRAs",
      "models": [
        {
          "name": "Ghibli Style LoRA",
          "file": "ghibli_style_lora.safetensors",
          "dest": "loras",
          "size_gb": 0.4,
          "status": "present",
          "on_disk_bytes": 429496729,
          "base_model": "WAN 2.1",
          "pair_id": "ghibli",
          "pair_role": "both",
          "tags": ["style", "anime"]
        }
      ]
    }
  ]
}

The response format is identical to GET /api/admin/models. Each LoRA entry may include pairing fields:

Field	Description
`pair_id`	Identifier grouping related LoRAs (e.g., high/low noise variants)
`pair_role`	`"high"`, `"low"`, or `"both"` -- the noise level variant
`base_model`	The base model this LoRA is compatible with

curl https://your-pod.runpod.io/api/admin/loras \
  -H "X-API-Key: your-api-key"

GET /api/admin/loras/compatible/{base_model}¶

List LoRA pairs compatible with a given base model. Used by the LoRA picker in workflow runner forms. LoRAs are grouped by pair_id.

Auth: Required

Path parameters:

Parameter	Description
`base_model`	The base model identifier (e.g., `"WAN 2.1"`, `"SDXL 1.0"`)

Response: 200 OK

[
  {
    "pair_id": "ghibli",
    "name": "Ghibli Style",
    "high": "ghibli_style_high.safetensors",
    "low": "ghibli_style_low.safetensors",
    "both": null
  },
  {
    "pair_id": "anime",
    "name": "Anime Style",
    "high": null,
    "low": null,
    "both": "anime_style_lora.safetensors"
  }
]

Field	Description
`pair_id`	Unique pair identifier
`name`	Clean display name (suffixes like "High Noise" are stripped)
`high`	Filename of the high-noise variant, or `null`
`low`	Filename of the low-noise variant, or `null`
`both`	Filename of the combined variant, or `null`

When both is set, the LoRA handles all noise levels. When high and low are set, the workflow should apply both with appropriate strength scheduling.

curl https://your-pod.runpod.io/api/admin/loras/compatible/WAN%202.1 \
  -H "X-API-Key: your-api-key"

POST /api/admin/loras/gallery/download-single/{image_id}¶

Download a single CivitAI image by its ID and add it to the gallery store. Used to add specific images without downloading an entire gallery.

Auth: Required

Path parameters:

Parameter	Description
`image_id`	CivitAI image ID

Response: 200 OK

{
  "status": "downloaded",
  "civitai_id": "12345678"
}

curl -X POST https://your-pod.runpod.io/api/admin/loras/gallery/download-single/12345678 \
  -H "X-API-Key: your-api-key"

POST /api/admin/loras/gallery/to-input/{image_id}¶

Copy a gallery image to the ComfyUI input directory so it can be used as an input image in workflows.

Auth: Required

Path parameters:

Parameter	Description
`image_id`	CivitAI image ID (from the gallery database)

Response: 200 OK

{
  "filename": "gallery_12345678.jpg"
}

The returned filename is the ComfyUI-relative path that can be used in workflow execute calls.

curl -X POST https://your-pod.runpod.io/api/admin/loras/gallery/to-input/12345678 \
  -H "X-API-Key: your-api-key"