ComfyUI Node
LocateAnything Grounding
Runs visual grounding frame by frame for an IMAGE batch. Task modes: - ground_multi: locate every matching instance. - ground_single: locate one matching instance. - detect: detect matching categories or descriptions. - ground_text: locate a requested text phrase. - detect_text: detect scene text; query is ignored. - gui_box: locate a GUI region. - gui_point: return a point for a GUI target. - point: return a point for a described target. - custom: use query as the complete prompt. Coordinates are parsed from the model's normalized [0, 1000] format. A batch of video frames is processed independently, with native ComfyUI progress updates.
LocateAnything Grounding
- model
- image
- answer
- locations_json
- annotated_image
- mask
- mask_overlay
◄taskground_multi►
◄queryperson►
◄generation_modehybrid►
◄max_new_tokens2048►
◄temperature0.00►
◄top_p0.90►
◄top_k0►
◄repetition_penalty1.10►
◄point_radius12►
◄mask_grow0►
◄mask_blur0.0►
◄overlay_color#00ff66►
◄overlay_opacity0.45►
◄seed0►
◄verbosetrue►
CategoryLocateAnything
Inputs (17)
| Name | Type | Default | Description |
|---|---|---|---|
| model | LOCATEANYTHING_MODEL | — | |
| image | IMAGE | — | |
| task | COMBO | ground_multi | Operation mode. Hover the node help (?) for the complete list. Use custom to send query as the full model prompt. |
| query | STRING | person | Description, text, GUI target, or full prompt for custom mode. |
| generation_mode | COMBO | hybrid | Decoding strategy: hybrid uses fast decoding with stable fallback; fast prioritizes speed; slow prioritizes the stable path. |
| max_new_tokens | INT | 20481–8192 | Maximum generated tokens. Reduce this when shorter answers are sufficient. |
| temperature | FLOAT | 0.000–2 | Sampling temperature. 0.0 is the safest path; the node retries with safe settings if higher temperatures fail. |
| top_p | FLOAT | 0.900–1 | Nucleus sampling cutoff. Relevant when temperature is above 0. |
| top_k | INT | 00–1024 | Top-k sampling cutoff. 0 disables top-k, matching the official worker. |
| repetition_penalty | FLOAT | 1.100–3 | Penalty for repeated tokens. The official worker uses 1.1. |
| point_radius | INT | 121–512 | Radius in pixels used to draw point results into the output mask. |
| mask_grow | INT | 0-512–512 | Grow the output mask by this many pixels. Negative values shrink it. |
| mask_blur | FLOAT | 0.00–100 | Gaussian blur radius applied after mask grow. Use 0 for hard edges. |
| overlay_color | STRING | #00ff66 | Hex RGB color used by the mask overlay preview, for example #00ff66 or #ff0000. |
| overlay_opacity | FLOAT | 0.450–1 | Opacity of the colored mask overlay preview. |
| seed | INT | 00–18446744073709550000 | Sampling seed. Relevant when temperature is above 0. For IMAGE batches, frame N uses seed + N. |
| verbose | BOOLEAN | true | Print the official generation step log in the terminal. Disable for quieter runs. |
Outputs (5)
| Name | Type | Description |
|---|---|---|
| answer | STRING | Raw model response. For batches, this is a JSON array of responses. |
| locations_json | STRING | Structured JSON with prompt, timing, normalized coordinates, pixel coordinates, and batch index. |
| annotated_image | IMAGE | Input image batch annotated with returned boxes and points. |
| mask | MASK | Post-processed mask batch after grow and blur: filled boxes and circles centered on returned points. |
| mask_overlay | IMAGE | Original image batch with the post-processed mask blended using overlay_color and overlay_opacity. |