Cindy Zhu.
← all free guides
Content creation Fish Audio

The Fish Audio voice-clone workflow

Hey, it's Cindy 🌱 Here's the full workflow from the reel: how to clone your own voice with Fish Audio and control the emotion right in the script. Every voice in that video, including the Mandarin one, was a clone, and this is exactly how it's done.
the setup

Clone your voice in a few minutes 🎙️

  1. 🐟 Make an account. Go to Fish Audio and make a free account (the free plan includes commercial use).
  2. 🎤 Open voice cloning. Choose Instant Voice Clone.
  3. ⬆️ Upload a clean sample. Upload a clean sample of you talking. It works with as little as 10 to 15 seconds, but around 30 seconds of clear, single-speaker audio gives better fidelity. No background noise, no music.
  4. ✨ Generate it. Generate the voice. Now you have a reusable clone you can type any script into.
the real trick

Control the emotion (the part that makes it sound human) 🎭

On the S2.1 model you tag the emotion right in the text, in square brackets, just before the line:

🎭 emotion tags in the script 
[excited] I cannot believe this actually worked.
[bored] I cannot believe this actually worked.
You can even layer them, like [sad][whispering] I miss you. There's a set of core tags (excited, angry, confident, whispering, shouting, laughing, and more), plus open-ended ones: S2.1 will read almost any feeling you write in brackets. This is the thing old AI voices could never do.
why bother

Why it's worth it 🌍

🗣️ 83 languages The free S2.1 Pro model (s2.1-pro-free) supports 83 languages, so you can voice content for an international audience in your own voice.
⚡ Built for speed It's built for low-latency streaming, multilingual text-to-speech, and voice cloning.
🆓 Free API window Fish Audio has a free developer window on the free S2.1 Pro model (model string s2.1-pro-free) right now (they've extended it before, so check the current status before you count on it). Great time to build.
🌱 The free plan 8,000 credits a month, up to 7 minutes, commercial use included. Paid plans start at $11/mo if you need more.
the craft

How to actually use it 💡

  1. 🎬 One clone, any script. Faceless content, ad voiceovers, bedtime stories in your own voice, or dubbing a reel into another language.
  2. 🔁 Keep a master sample. Keep one clean master sample of your voice so every clone sounds consistent.
  3. ✍️ Tag as you write. Write the emotion tags as you write the script, not after. The delivery changes the whole feel.
the payoff

10 ways to make money with AI voices 💸

Once you can generate unlimited audio in your own voice, that voice becomes an asset you can rent out again and again, without ever sitting in front of a mic. Here are ten real ways people turn a cloned voice into income:

  1. Faceless video channels. Narrate faceless YouTube, TikTok and Reels in your voice, and earn from ad revenue, sponsors and affiliate links.
  2. Voiceover gigs. Sell voiceovers on Fiverr and Upwork and deliver in minutes, no booth and no re-records.
  3. Audiobook narration. Narrate your own ebooks, or other authors', for audiobook platforms, or offer it as a done-for-you service.
  4. Ad and UGC reads. Spin up dozens of ad-script variations for brands and performance marketers who are testing hooks at volume.
  5. Dubbing and localisation. Voice the same content in other languages (S2.1 handles 83) and sell it to creators going global.
  6. Bedtime, sleep and meditation apps. Subscription apps that read aloud in a warm voice (that's the very next guide).
  7. Course and e-learning narration. Narrate lessons and modules for course creators who dread recording themselves.
  8. Local business audio. Explainer videos, phone greetings and on-hold messages for small businesses near you.
  9. Podcast production. Intros, outros, sponsor reads, or an entire show narrated start to finish.
  10. Rent your voice out. Sell a "voice pack," or set up a founder's cloned voice so it can front all of their content for them.
🌱 The honest version: none of these pay just because you own a clone. They pay when you pair the voice with a real audience, a real service, or a real product. The clone removes the recording bottleneck, you still bring the distribution.

Clone your voice here: Fish Audio 🐟

Follow @cindiezhu for more AI tips every single day 🌱