Get Image From Prompt (Fallback)

Look at this node's inputs and outputs and you'll find them identical, field for field, to SDWebuiAPINode: same prompt/api_endpoint required, same optional auth, seed, negative_prompt, steps, width, height, hr_scale, hr_upscale, enable_hr, cfg_scale, same IMAGE output. Display name "Get Image From Prompt (Fallback)" is the tell - it's positioned as the fallback implementation of the exact same feature, not a different feature. The pack's README doesn't spell out the internal difference, but "fallback" alongside an identical schema strongly suggests it's built on a different HTTP path under the hood (a lighter-weight or lower-dependency client, most likely) so that if the primary node's request library isn't available or is having trouble, this one is still usable without changing anything else about how you use it.

Both nodes live in aria1th's ComfyUI-LogicUtils - a personal utility pack from the same handle behind the Illustrious XL training work, bridging ComfyUI to any Automatic1111-compatible WebUI's REST API (A1111 itself is frozen since mid-2024, but its /sdapi/v1/* contract lives on in active forks like Forge and reForge, so this bridge still has real use if you're running one of those alongside ComfyUI).

How it works

Functionally the same as SDWebuiAPINode: it POSTs your prompt and parameters to the target WebUI's txt2img API endpoint, waits for the render, and returns the resulting IMAGE. The practical way to use having both nodes: if the primary node throws an import or connection error you can't quickly resolve, drop this one in as a drop-in swap - the inputs line up exactly, so no rewiring needed beyond the one connection.

Inputs and outputs

Required:

• prompt (STRING) - the positive prompt.
• api_endpoint (STRING) - the target WebUI's API base URL, e.g. http://127.0.0.1:7860.

Optional: negative_prompt, seed (default -1), steps (default 28), width/height (default 1024×1024), cfg_scale (default 7), enable_hr (default off) with hr_scale (default 1.5) and hr_upscale (default "Latent"), and auth for basic-auth credentials.

• Output: IMAGE - the rendered image from the remote WebUI.

How to install it

ComfyUI Manager: search "ComfyUI-LogicUtils", install, restart - this installs both API nodes together, there's no separate install path for the fallback variant. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/aria1th/ComfyUI-LogicUtils

then restart ComfyUI.

Common issues & troubleshooting

Trying this node because the primary one errored, and it errors too. If both fail the same way, the problem is almost certainly the target WebUI (not running, API disabled, wrong endpoint URL) rather than either node - the fallback shares the same target, it just may reach it differently.

Import error on ComfyUI startup. Same as the primary node: the pack's dependency auto-install is opt-in per the README. Set COMFYUI_LOGICUTILS_AUTO_INSTALL=1 before launch to let it install what's needed, or COMFYUI_LOGICUTILS_SKIP_INSTALL=1 if the auto-install hook itself is the issue.

Connection refused / timeout. The remote WebUI needs --api enabled and needs to actually be reachable - same networking checklist as the primary node: same host/port, no firewall blocking it, container networking configured if ComfyUI and the WebUI are in separate containers.

Not sure which of the two nodes to use by default. Reach for the plain SDWebuiAPINode first; keep this one in your back pocket for the specific case where the primary one won't load or won't connect and you want to try an alternate path without redoing your whole graph.