Fish Audio: Text to Speech
Generate natural, expressive speech from text with Fish Audio S2 Pro, a text-to-speech model supporting 80+ languages and fine-grained control over emotion, tone, and pacing using inline tags like [whisper] or [excited]. Input your text and receive one high-quality audio output, with optional style tags woven directly into the script. Ideal for voiceover production, multilingual content creation, and character dialogue in games or animations.
About
This ComfyUI workflow turns written scripts into natural, expressive speech using the Fish Audio S2 Pro text‑to‑speech model. It supports over 80 languages and lets you direct performance with inline style tags (for example, adding [whisper] or [excited] directly in your text). The core path runs your text through the FishAudioTextToSpeech node and saves a single high‑quality audio file via SaveAudioAdvanced.
What you provide
- Reading the workflow
Runs in ComfyUI. This workflow can't run on this page. Open it in Comfy Cloud with everything loaded, or download it for your own ComfyUI.
Try in CloudQuestions & answers
Place a tag like [whisper] or [excited] directly in your script before the words you want to affect. The FishAudioTextToSpeech node forwards your text to the Fish Audio S2 Pro API, which interprets these tags to adjust tone, emotion, and pacing. For the full list of supported tags and syntax, consult the Fish Audio documentation.
Preset voices (FishAudioVoiceSelector) are curated, consistent, and ready to use. Instant cloning (LoadAudio + FishAudioInstantVoiceClone) adapts the TTS to a short reference clip so you can match a specific speaker. Cloning quality depends on the reference audio—clean speech with minimal noise typically yields better results.
Use 5–30 seconds of clear, solo speech with no music or heavy reverb. A dry, well‑recorded segment where the speaker talks at a natural pace works best. Load this file in the LoadAudio node, then run FishAudioInstantVoiceClone to generate the voice profile for the TTS node.
Yes. Fish Audio API nodes are pay‑per‑use and require you to be signed in. Usage, quotas, and pricing are managed by your account with the service. If requests fail, check your authentication status, balance, and any rate limits in your account.