OTX Versatile Multiple Inputs 4

The pitch of OTX Versatile Multiple Inputs 4 is a single box where each of four slots can hold an int, a float, or a string - you type the value, pick the type from a dropdown, and the node hands it back as whatever you chose. Four different data types from one node instead of four single-type nodes. It's part of Otonx's Custom Nodes, a small personal pack by Budi Hartono that gets essentially zero community discussion, and it's worth reading this one with your skeptic hat on, because the type dropdown has a real bug hiding in the source.

How it works (and where it breaks)

Each slot is a multiline textbox plus a valuetype_N dropdown offering int, float, or string. In the code, the conversion is straightforward:

if valuetype_1 == "INT":
    value_1 = int(value_1)
elif valuetype_1 == "FLOAT":
    value_1 = float(value_1)
elif valuetype_1 == "STRING":
    value_1 = str(value_1)

Here's the thing: the dropdown emits lowercase int/float/string, but the checks compare against uppercase "INT"/"FLOAT"/"STRING". Those never match. So that entire block never fires, and every slot returns the raw text you typed, as a string - no matter what you set the dropdown to. It's a one-word-per-branch case mismatch, and it means the node currently doesn't convert anything.

The outputs are wildcard (*), so ComfyUI lets you wire them into anything. That's the trap: the connection succeeds, and then the downstream node gets a string where it asked for a number, which can error or behave oddly a few steps later. If you're typing "512" into a slot and feeding it to something that wants an INT, don't trust the dropdown - the number isn't a number until the consumer converts it.

The inputs and outputs that matter

Four slots, each with:

• value_N - multiline textbox, default empty, where you type the value
• valuetype_N - dropdown: int / float / string

Four outputs (value_1 through value_4), all wildcard type. The genuinely good use of this node today is text: the multiline textbox makes it a decent place to park prompts or JSON snippets in one visible block. For numbers, treat the dropdown as decoration and either convert downstream or stick with a proper typed node.

Installing it

The pack is pure Python, no dependencies, no model downloads. Install via ComfyUI Manager (search "Otonx") or clone:

cd ComfyUI/custom_nodes
git clone https://github.com/budihartono/comfyui_otonx_nodes

Restart ComfyUI and find it under OtonxPack.

Common issues

• Numbers come out as text - the case mismatch above. Verify by wiring the output into a Show Text / debug node and checking the type.
• A node two hops downstream fails on a "string" where it expected an INT - that's this bug surfacing late. Convert explicitly, or use the Integer Multiple Inputs node for numbers.

A 0-impression utility node with a conversion bug is not a hill worth dying on - but knowing the dropdown is cosmetic saves you a genuinely confusing hour of debugging.