Conditioning Grid (string)
Type prompts straight into the grid, skip the encoders
- clip
- CONDITIONING
The cond version of Conditioning Grid is great, but you still have to CLIP-encode every cell's prompt before it can go in. Conditioning Grid (string) removes that step entirely: each cell takes a plain text string, the node encodes it with your CLIP model internally, and you get the same grid-regional conditioning out. For a quick multi-prompt layout, this is the version you'll actually type into.
How it works
Mechanically it's the string convenience wrapper around Conditioning Grid (cond): it runs CLIPTextEncode on each cell string and on the base string, then hands those conditionings to the exact same grid-assembly logic (ConditioningSetArea per cell + ConditioningCombine). So you're not losing any capability by using the string version for text-driven grids - the encoding is just happening inside the node.
Inputs:
- clip - your CLIP model, required.
- base - a string for the shared base conditioning. Empty string is the default and the sensible starting point; put shared scene elements here if you want them across all cells.
- columns / rows - grid size (1–8, default 3×3).
- width / height - cell size in pixels (default 256).
- strength - applied to every cell (default 3; lower it if regions bleed).
- The cell text inputs -
r1_c1,r1_c2, ... as plain STRING sockets. Same drill: right-click → "update inputs" to generate them for your chosen grid size.
Output is a single CONDITIONING ready for the KSampler.
When to pick string vs cond
String version wins when your cells are all plain prompts - the common case. It saves you the encoder row entirely. The cond version is for when cells need pre-built conditionings: a CLIPEncodeMultiple output, a ControlNet-applied conditioning, a pooled/multimodal embedding, or anything that isn't just a CLIPTextEncode string. If every cell is just a sentence, use this node and don't look back.
Two things to know going in:
- Every cell is encoded with the same CLIP, so no per-cell CLIP choice.
strengthis shared across all cells, same as the cond version. Per-cell strength still means the hand-built route.
Install
cd ComfyUI/custom_nodes
git clone https://github.com/bmad4ever/comfyui_bmad_nodes
cd comfyui_bmad_nodes
pip install -r requirements.txt
or install "comfyui_bmad_nodes" via ComfyUI Manager and restart. No extra dependencies beyond the pack's shared set. It's the least-effort way to do regional prompts in this pack, and honestly one of the pack's best reasons to exist.
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| clip | CLIP | — | |
| base | STRING | — | |
| columns | INT | 31–8 | — |
| rows | INT | 31–8 | — |
| width | INT | 25616–2048 | — |
| height | INT | 25616–2048 | — |
| strength | FLOAT | 3.00 | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| CONDITIONING | CONDITIONING | — |