Tutorial2026-03-13·8 min read

How to Use Fish Audio API for Audiobook Production

Use Fish Audio models on Kitta AI to create audiobook narration. This guide uses the Kitta AI workspace and API; accounts, API keys, credits and pricing belong to this platform and are separate from the model provider's services.

Why Fish Audio for Audiobooks?

Fish Audio's S1 model ranked #1 on TTS-Arena2, known for emotion control and naturalness. For audiobook production, it has three key advantages:

Voice Cloning

Create a licensed narration voice from a short audio sample and maintain consistent narration throughout the book.

Emotion Control

Open-domain emotion annotation makes dialogue scenes vivid, not robotic.

Fish Audio S2.1 Pro: 83 Languages

Train once, publish in multiple languages — ideal for international audiobook releases.

Production Workflow: 5 Steps

1

Prepare reference audio

Record or collect 10–30 seconds of clean audio. The cleaner the sample (no background noise), the better the clone. Supports MP3, WAV, M4A.

2

Create a voice model in Kitta AI

Log in to Kitta AI, go to Voice Cloning, upload your reference audio, name the voice, and click Start Cloning. Training completes in about 1 minute.

3

Split your manuscript

Divide the manuscript by chapter. Keep each segment under 1,000 words. Use Kitta AI's Long Text Mode or Batch Mode — the system handles splitting automatically.

4

Generate and download audio

Select your cloned voice model, paste the text, and click Generate. Batch-generate multiple chapters, then download the MP3 files.

5

Post-processing (optional)

Use Audacity or Adobe Audition to normalize volume across chapters and stitch them together for the final audiobook file.

Tips for Better Quality

Include emotional variation in your reference audio (not just flat reading) — the cloned voice will be more expressive
For dialogue, add emotion cues in the text like "(excitedly)" — Fish Audio supports natural language emotion control
Keep each text segment to 500–800 words for best quality on long-form content
Use the same voice model for all chapters of a book to maintain consistency

FAQ

Is Fish Audio API good for audiobook production?

Use Fish Audio S2.1 Pro through the Kitta AI API for authorized voice cloning and audiobook narration in 83 languages. Reference-audio requirements, text limits and streaming options depend on the selected model. Use Kitta AI API keys and credits, not the model provider's account.

How much does it cost to produce an audiobook with Fish Audio?

Fish Audio offers a free tier (8,000 credits/month, ~7 minutes of audio). Via Kitta AI, which uses Fish Audio technology, the free plan includes 1,000 credits and paid plans start at 20,000 credits/month.

How many credits does a 100,000-word audiobook require?

At standard Kitta AI pricing, 100,000 characters requires approximately 100,000 credits (1 credit = 1 character). The Pro plan includes 20,000 credits/month — for a full audiobook, a top-up plan is recommended.

Start Your Audiobook Today

Kitta AI is powered by Fish Audio technology. Try voice cloning and audiobook production for free.

Related