Kling Voice Catalog

The pack's TTS nodes default to a voice called girlfriend_4_speech02, and if that's all you ever hear you'll assume Kling's voices are mediocre. They're not - you're just hearing the starter voice. KlingDirect_VoiceCatalog is the node that shows you the whole menu: it outputs the full preset voice catalog as a JSON string, so you can actually see what voices exist and grab the voice_id you want before wiring up speech.

It's the simplest node in the pack - zero inputs. Wire it up, run it, and read the catalog_json output. The JSON is a list of entries, each with a display_name, the voice_id you paste into Kling TTS / Kling TTS Advanced / Kling Lip Sync's voice_id field, and the language the voice speaks. It's the human-readable companion to the Voice Selector node's dropdown - that node gives you the same presets as a picker widget; this one gives you the raw data, including every ID at once.

When you'd use it

• Choosing a voice deliberately. Read the catalog, pick a display_name that suits the character, and paste its voice_id where TTS wants one.
• Finding voices for a specific language. The language field per entry lets you filter for the right voice instead of testing blind.
• Driving voice choice from data. Since the output is a STRING of JSON, you can feed it to a JSON-parsing node and even select a voice programmatically in a bigger workflow.

Output

One output, catalog_json (STRING) - pretty-printed JSON of the full preset list. No auth required, no API call, no credits; the catalog is bundled with the pack itself, not fetched from Kling. That last point is worth noting: because it's a local constant, it reflects the pack's bundled list rather than the live server state. If Kling adds new voices upstream, you'd see them after the pack updates, not instantly.

Install

Part of ComfyUI-Kling-Direct: ComfyUI Manager → search "Kling Direct" → install, or git clone https://github.com/IxMxAMAR/ComfyUI-Kling-Direct into ComfyUI/custom_nodes, then restart. No key, no dependencies beyond ComfyUI's built-ins - this node is pure local config, which makes it a safe first node to run when you're just poking around the pack.

The gotcha

The default girlfriend_4_speech02 that TTS ships with is in the catalog - it's a real voice, just a generic one. The catalog exists precisely so you stop using it. And if you've cloned a voice with the Voice Clone node, that custom voice_id won't appear here (it's not a preset) - keep it from the clone node's output instead.

For everything else: run this once, keep the JSON handy, and your next TTS call will actually sound like you chose it.