Comflowy Omost Preview

The Omost workflow is all about layout, and layout is exactly what you can't see from a pile of JSON. Comflowy Omost Preview renders the canvas for you: you feed it an OMOST_CANVAS_CONDITIONING object and it draws what the regions actually look like - where the "left" box sits, how big the "center" area is, how the global description fills the frame. One output is the rendered image; the other is the underlying canvas JSON, so you get both the picture and the data.

Why this node matters: Omost's area conditioning is powerful precisely because it's explicit, but it's also easy to get wrong in ways you only discover after a full sampler run. The whole point of this node is to make that feedback loop cheap. Wire Comflowy Omost LLM into it and you can eyeball the LLM's layout before it ever reaches a KSampler - catch the "everything is crammed into the left third" failure in seconds instead of after a minute of wasted generation. It's the same role a preview node plays anywhere in ComfyUI: see before you commit.

How it works and the inputs

The node renders the canvas to an initial latent and converts it to an image using the pack's Omost canvas machinery - effectively a quick visual mockup of the region masks and colors, not a real diffusion render. It's fast and it's free (no API, no credits).

• canvas_conds (required) - any OMOST_CANVAS_CONDITIONING source: Comflowy Omost LLM, either of the two Load Canvas nodes, or anything else in the pack that emits the type.

Outputs:

• IMAGE - the rendered layout preview.
• JSON - the canvas condition data itself, so you can inspect or pass it on.

Installing and the gotchas

Install Comflowy's Custom Nodes via ComfyUI Manager (search "Comflowy") or git clone https://github.com/6174/comflowy-nodes into custom_nodes, restart. No API key, no downloads - this one is pure local rendering.

A couple of honest notes:

• It's a layout preview, not a quality preview. The rendered image shows region placement and color coding; it tells you nothing about how the final image will look aesthetically. Use it to validate composition, not to judge art.
• The JSON output is the same data you already had. It's there for inspection and for passing into other nodes - don't expect new information beyond the rendered layout.
• Feed it from anywhere in the Omost chain. If you edit a canvas via the Load Canvas nodes, run the preview again to sanity-check your edits before the conditioning step.

If you're building any Omost-style workflow in this pack, this node deserves a permanent place in the graph - one tiny node that turns "I wonder what that layout actually is" into a two-second answer.