TJ Studio ONE — Text Output

TJ Studio ONE - Text Output is the most anonymous-looking node in the TJ NODE STUDIO ONE pack, and it's also the one you'll almost never add by hand. It's not a generator and it doesn't process anything. It exists to solve a plumbing problem inside the pack, specifically inside the MiniMax H3 node's Image → Brief feature.

Here's the situation, from the node's own source comment. When you use Image → Brief in MiniMax H3 ONE STUDIO with the native vision path, ComfyUI submits a small side graph that ends in a TextGenerate node. TextGenerate produces a STRING, but nothing in that graph consumes it, and TextGenerate itself doesn't emit a UI result - so the generated brief has nowhere to go. TJ Studio ONE Text Output fixes exactly that: you wire the string into its single input, and because it's marked as an output node, the string becomes a ui.text payload that the same websocket machinery the video graph already uses can pick up and display. The frontend just reads byNode[thisNodeId].text[0] and shows you the brief.

In other words: it's a terminal node. Give it a string, and it makes sure that string is visible instead of vanishing into the void between queue submissions.

Inputs and behavior

The whole schema is one required field:

• text (STRING, force-input) - the string you want to surface. That's it.

There are no outputs, by design. It's an output node, so it sits at the end of a graph, catches a string, and reports it to the UI. If you're poking at it directly, connect a text/string-producing node into text and run - you'll see the value appear in the node's result. Nothing gets written to disk, no image comes out, no file path comes back. It's a display of what's already in the graph.

When you'll actually meet it

Realistically, two situations. Either you're reading a shared workflow from someone using MiniMax H3's native Image → Brief flow and this node is sitting there doing its quiet job, or you're the kind of person building your own side graphs and you need a terminal string target with a UI payload. Both are fine. You do not need to install anything extra for it - it ships inside the same pack as the MiniMax node, which is installed the same way:

cd ComfyUI/custom_nodes
git clone https://github.com/designloves2/ComfyUI-TJ_NODE_STUDIO_ONE.git

or via ComfyUI Manager by searching ComfyUI-TJ_NODE_STUDIO_ONE, then restart.

Worth saying plainly: if you never touch MiniMax H3's Image → Brief, you'll probably never have a reason to care about this node. That's fine - it's a piece of internal scaffolding the pack exposes because ComfyUI needs real nodes to carry UI results, and hiding it would have meant the native vision pipeline couldn't show you what it wrote. Don't expect much from it, and it won't disappoint.