Images Input ♾️Mixlab

Worth clearing up first: this isn't one of Mixlab's Workflow-to-APP form controls. The README's list of node types that become app inputs is specific - Load Image, VHS_LoadVideo, CLIPTextEncode, PromptSlide, TextInput_, Color, FloatSlider, IntNumber, CheckpointLoaderSimple, LoraLoader - and Images Input isn't on it. What it actually does is narrower and more useful for a different situation: you already have an image as base64 text, and you want it decoded into a real IMAGE without going through a file-upload widget at all.

That situation comes up more than you'd think - calling a workflow via API, receiving an image inside a JSON payload, or just pasting base64 image data straight into a graph for testing. Base64 is the standard way an image travels inside JSON, and this node is the decode step.

How it works

You provide image_base64 (the raw encoded string) and a text field alongside it. The node decodes the base64 into a normal IMAGE tensor, and passes text straight through unchanged - it's not processed, just carried alongside the image as a paired caption or prompt.

The inputs and outputs that matter

• image_base64 (STRING, default empty, required) - the base64-encoded image data.
• text (STRING, multiline, default empty, required) - a caption or prompt to travel alongside the image.

Two outputs: image (IMAGE) - the decoded picture - and text (STRING) - the passthrough field, unchanged.

How to install it

Ships with the Mixlab pack.

• ComfyUI Manager: search comfyui-mixlab-nodes, install, restart.
• Manual: cd ComfyUI/custom_nodes && git clone https://github.com/shadowcz007/comfyui-mixlab-nodes.git, install requirements (install.bat on Windows, or ../../../python_embeded/python.exe -s -m pip install -r requirements.txt), restart.

No model downloads for this node - decoding base64 is pure Python.

Common issues & troubleshooting

Decoding fails outright. If you're pasting from a data URL (data:image/png;base64,...), strip the data:image/...;base64, prefix before feeding it into image_base64 - this field wants the raw payload, not the full data URL. This node has no reason to validate malformed or truncated base64 gracefully, so garbage input is more likely to error than silently produce a broken image.

Expecting this to appear as an app form field. It won't - Images Input isn't on the README's supported list of Workflow-to-APP input node types, so publishing a workflow with this node in it won't turn it into an upload widget on the app page. If you want an app-facing image upload, use the pack's Load Image node instead, or a plain ComfyUI Load Image node, which is on that list.

General install trouble. This pack bundles a lot of unrelated functionality (LLM chat, TTS, screen sharing, 3D) behind one shared requirements.txt - if setup errors on something unrelated to image handling, run the manual pip install -r requirements.txt and check which specific package is actually failing before assuming the whole install is broken.