builders · review · first seen Oct 27, 09:33 UTC

ComfyUI adds image and partner nodes in v0.43.0

ComfyUI v0.43.0 shipped a Pad Image node, HeyGen Voice partner nodes, and 2K/4K options for FLUX 3 Video partner nodes. Builders should care because these additions may change where image preparation, voice generation, and video resolution selection sit in a workflow.

L0 · primary1 source3 claims checked4 min read

What shipped

The v0.43.0 release adds a Pad Image node with mirror padding and pad to multiple. Those are the two controls established by the release notes. Mirror padding is worth testing when a workflow needs space around an image; pad to multiple is worth testing when a later step expects dimensions aligned to a multiple. The source does not specify the node’s inputs, its defaults, or the resulting dimensions for a particular image, so inspect the output rather than assuming either control produces the shape you need. [1]

The release adds HeyGen Voice instant clone and text-to-speech partner nodes. This establishes that the nodes shipped in ComfyUI v0.43.0, not that a particular voice job will run without setup or produce an acceptable result. If voice is part of an existing workflow, the immediate question is whether these nodes fit the inputs and outputs you need. The excerpt gives no setup sequence, supported languages, usage terms, pricing, or output limits. Keep those as checks before making the nodes a required step. [1]

The FLUX 3 Video partner nodes gained 2K/4K resolution options in v0.43.0. This is a choice added to those nodes; the release excerpt does not say that every workflow can select it or that a generated result will meet a particular quality target. It also does not establish generation time, file size, price, or hardware requirements. Treat the new options as settings to test on a representative job, not as evidence that a higher-resolution workflow is ready to replace the one you use now. [1]

Why it matters if you build

For an image workflow, the useful change is a place to try padding inside ComfyUI before passing an image to the next step. A controlled comparison would use the same input, apply each Pad Image control separately, and inspect both the added area and final dimensions. That would tell you whether the node removes a preparation step in your own workflow. The release does not provide a compatibility list for downstream nodes, so a padded image still needs to pass through the rest of your workflow before you rely on it. [1]

For workflows that combine media steps, the partner-node additions create two distinct decisions. HeyGen Voice nodes offer instant clone and text-to-speech paths; FLUX 3 Video nodes offer 2K/4K choices. Neither addition establishes an end-to-end voice-and-video workflow, and the excerpt supplies no costs or limits for either partner service. Test each addition on its own before joining them in a workflow. Otherwise, a failure at sign-in, generation, or handoff could be hard to attribute from these release notes alone. [1][2]

How to try it today

1. Open a workflow in ComfyUI v0.43.0 and locate the Pad Image node.
2. Run one image through mirror padding, then through pad to multiple; inspect each output and its dimensions.
3. Locate the HeyGen Voice instant clone and text-to-speech partner nodes; check what setup and inputs they request before depending on either.
4. Locate the 2K and 4K choices in a FLUX 3 Video partner node; test a representative job only if your setup permits it.
5. Keep the previous workflow path available until the tested outputs fit the next steps you actually use.

What we checked

Every claim below was walked down the source ladder. The rung is how close the source is to the thing itself; the tense is the source's own.

1. shipped L0 · primary

   ComfyUI v0.43.0 adds a Pad Image node with mirror padding and pad to multiple.

   per GitHub releases · comfyanonymous/ComfyUI (L0)
2. shipped L0 · primary

   ComfyUI v0.43.0 adds HeyGen Voice instant clone and text-to-speech partner nodes.

   per GitHub releases · comfyanonymous/ComfyUI (L0)
3. shipped L0 · primary

   ComfyUI v0.43.0 adds 2K/4K resolutions to FLUX 3 Video partner nodes.

   per GitHub releases · comfyanonymous/ComfyUI (L0)

What is not confirmed

These release notes do not establish Pad Image defaults or downstream compatibility. They do not give HeyGen Voice setup requirements, pricing, usage terms, or output limits. They do not give FLUX 3 Video 2K/4K pricing, generation time, file size, hardware requirements, or a quality comparison. Do not use the existence of a node or resolution option as a substitute for those checks. [1][2][3]

Who should act, who should wait

Move now if padding, voice, or a higher video resolution is already a concrete requirement: v0.43.0 gives you specific nodes or options to evaluate against a real input. Wait before making any of them mandatory in a production workflow if your decision depends on costs, limits, access, or downstream behavior. The release confirms what was added, but it does not answer those operational questions. A small comparison today can establish whether the new path works for you without assuming that shipping a control settles the adoption decision. [1][2][3]

Sources

• v0.43.0 — GitHub releases · comfyanonymous/ComfyUI L0 · primary