ComfyUI-HMT-Ultimate Introduction
ComfyUI-HMT-Ultimate adds custom nodes to ComfyUI for turning Vietnamese text into speech. You can choose a ready-made voice for a quick voiceover or provide a short recording as a reference for a new spoken line. A third node shows which preset voices are available.
For an AI art workflow, these nodes can help you add narration, dialogue, or spoken introductions to a project without leaving ComfyUI. Both speech-generating nodes provide audio for use in your workflow and the path to a saved audio file.
How ComfyUI-HMT-Ultimate Works
Think of the extension as giving you two ways to cast a voice for your text:
- Choose a preset voice: Write what you want spoken, then select a voice from the node’s dropdown. The node generates speech using that selection.
- Use a reference recording: Provide a short audio sample and the words spoken in it. The voice-clone node uses those together as a reference when speaking your new text.
In either case, you enter the words you want to hear and run the ComfyUI workflow. The speech node returns audio data and a filepath pointing to the saved audio file. The filename_prefix setting lets you choose the beginning of that file’s name, which can make multiple takes easier to identify.
ComfyUI-HMT-Ultimate Features
VieNeu TTS (Preset Voice)
Use this node when you want to generate Vietnamese speech without preparing a reference recording.
| Input | What it does |
|---|---|
text | The Vietnamese words you want spoken. |
voice | Selects a voice from the available presets. |
filename_prefix | Sets the beginning of the saved audio file’s name. |
The node outputs audio and filepath. To try it, add VieNeu TTS (Preset Voice), enter a line such as “Xin chào, tôi là VieNeu TTS,” choose a voice from the dropdown, and run the workflow. You can then try another preset with the same text to compare the voices.
VieNeu TTS (Voice Clone)
Use this node when you have a short voice recording you want to use as a reference for new speech.
| Input | What it does |
|---|---|
text | The new words you want spoken. |
ref_audio_path | The path to a reference audio file. The author specifies a sample of 3–5 seconds. |
ref_text | A transcript of the words spoken in the reference audio. |
filename_prefix | Sets the beginning of the saved audio file’s name. |
The node outputs audio and filepath. Prepare the reference recording and its matching transcript, enter their details in the node, add your new text, and run the workflow. Make sure ref_text describes the reference recording, not the new line in text.
VieNeu List Voices
This node provides voices_list, a list of the available preset voices. Use it when you want to see your voice options before choosing one for VieNeu TTS (Preset Voice). The provided documentation does not name the individual presets, so the node is the source for the voices available in your setup.
Troubleshooting ComfyUI-HMT-Ultimate
If you do not see the nodes in ComfyUI, restart ComfyUI and check again; the author lists restarting as the final setup step.
If a voice-clone result is not what you expected, check the inputs before running another take:
- Confirm that
ref_audio_pathpoints to the recording you intended to use. - Check that the reference recording is approximately 3–5 seconds long.
- Confirm that
ref_textmatches the words spoken in that recording. - Check that
textcontains the new words you want generated.
If you are looking for a saved result, check the node’s filepath output. The audio output is the audio data available to the workflow; filepath identifies the saved file.
Frequently asked questions
- Do I need a reference recording to use the extension? No. VieNeu TTS (Preset Voice) uses a voice you select from the presets. A recording is needed for VieNeu TTS (Voice Clone).
- Where can I see the preset voices? Use VieNeu List Voices, which outputs
voices_list. - What should I put in
ref_text? Enter the words spoken in your reference audio, not the new words you want the node to say. - How can I distinguish saved takes? Change
filename_prefixto give each output a recognizable beginning to its file name.
Learn More about ComfyUI-HMT-Ultimate
- VieNeu-TTS SDK: The speech technology credited by the author.
- ComfyUI: The application in which these custom nodes run.
The original documentation identifies the extension’s license as the MIT License.
