ComfyUI Node
Danbooru Image Search
A ComfyUI node in Danbooru with 19 inputs and 6 outputs.
Danbooru Image Search
- positive_image
- negative_image
- IMAGE
- URLS
- IDS
- TAG
- IMG_WIDTH
- IMG_HEIGHT
◄download_sizepreview►
◄max_images4►
◄positive_prompt►
◄negative_prompt►
◄api_username►
◄api_key►
◄model_typeclip►
◄tag_embedding_modeauto►
◄combine_strategyBalanced►
◄tag_new_info_onlyfalse►
◄reduce_negative_overlapfalse►
◄idf_weight_tagsfalse►
◄fetch_modeauto►
◄extra_api_search0►
◄debugfalse►
◄image_weight1.00►
◄tag_weight1.00►
CategoryDanbooru
Inputs (19)
| Name | Type | Default | Description |
|---|---|---|---|
| download_size | COMBO | preview | Image size (preview -> thumbnail -> large -> full) attempts progressively larger variants if available. |
| max_images | INT | 41–50 | — |
| positive_imageopt | IMAGE | — | |
| negative_imageopt | IMAGE | — | |
| positive_promptopt | STRING | Comma-separated tags (e.g. 'touhou, reimu_hakurei') or text description to search for | |
| negative_promptopt | STRING | Comma-separated tags (e.g. 'monochrome, sketch') or text description to avoid | |
| api_usernameopt | STRING | — | |
| api_keyopt | STRING | — | |
| model_typeopt | COMBO | clip | Text/image embedding backbone to use (requires corresponding *_text.onnx for tags). |
| tag_embedding_modeopt | COMBO | auto | How to build tag embeddings: auto (prefer ONNX with fallback), hash_only (deterministic 1024D), onnx_only (real tags only). |
| combine_strategyopt | COMBO | Balanced | Balanced: normalized side sums diff; Stepwise: sequential de-overlap (orthogonalized sequence). |
| tag_new_info_onlyopt | BOOLEAN | false | Use only tag information that is not already present in the image (removes tag components that duplicate the image). |
| reduce_negative_overlapopt | BOOLEAN | false | Make negatives ignore what already matches positives (remove negative components that overlap with the positive mix). |
| idf_weight_tagsopt | BOOLEAN | false | Down-weight very common tags and up-weight rare ones using counts from selected_tags.csv. |
| fetch_modeopt | COMBO | auto | How to fetch metadata: auto (API then HTML fallback), api only, or html only. |
| extra_api_searchopt | INT | 00–1000 | Additional recovery attempts beyond top count if some posts fail. |
| debugopt | BOOLEAN | false | Extra debug logging. |
| image_weightopt | FLOAT | 1.000–4 | Weight of image embedding in combination. |
| tag_weightopt | FLOAT | 1.000–4 | Weight of tag embedding in combination. |
Outputs (6)
| Name | Type | Description |
|---|---|---|
| IMAGE | IMAGE | — |
| URLS | STRING | — |
| IDS | STRING | — |
| TAG | STRING | — |
| IMG_WIDTH | INT | — |
| IMG_HEIGHT | INT | — |