Clone Your Voice and Add Real Emotion (Angry, Sad, Whisper on Command)
Most AI voices sound flat because people type words and forget the feeling. This is how I clone a voice in under a minute and then direct it, angry, sad, whispering, embarrassed, just by dropping a tag in the text. Clone your own voice, or one you have clear permission to use. Do not impersonate real people, that is where this stops being fun and starts being illegal.
Clone your voice in under a minute
Open Fish Audio and go to the voice cloning tool. You record about 10 to 30 seconds of clean audio, or upload a clip you already have. Name the voice, save it, and it is ready to use in text to speech. No studio, no training wait, you are talking as yourself in about a minute.
The cleaner your sample, the better the clone. Record somewhere quiet, talk normally, and give it a couple of full sentences with a bit of range rather than one flat line.
The trick nobody uses: emotion tags
Here is what separates a robotic clone from one that fools people. You do not type flat words, you tag the emotion right in the text with square brackets. Whatever sits inside the brackets, the model performs. Put the tag right before the words it should color, exactly like this:
Every word you just heard was [emphasis] NOT me
Fish Audio gives you two sets of tags:
Emotional tone: [angry] [sad] [embarrassed] [emphasis] [whispering] [soft] [breathy] [excited]
Audio effects: [laughing] [chuckling] [moaning] [clear throat] [sobbing] [crying loudly] [sighing] [panting] [groaning] [crowd laughing] [background laughter] [audience laughing] [pause] [long pause]
Mix tone and effects freely, and switch as many times as you want inside one script.
My exact tag script
Paste this in, then swap in your own words. Notice the tag sits right before the line it changes.
Every word you just heard was [emphasis] NOT me.
[whispering] I need to tell you something.
[angry] You really think you can walk away from this?
[sad] I didn't mean for it to end like this. [sighing]
[excited] This is the best day of my life!
[embarrassed] I can't believe I said that out loud. [chuckling]
[soft] Just breathe. [long pause] Okay, I'm ready.Skip the manual work: hit Auto Tag
If you do not want to place tags by hand, use the Auto Tag button at the top of the editor. Paste your plain script, tap Auto Tag, and the AI reads the meaning of your words and drops the right tone and effect tags in for you. It is scary accurate, a scared line comes out scared, a punchline lands with a [chuckling], all without you marking a thing.
My workflow: hit Auto Tag to lay down a first pass on a long script, then hand place a tag or two on the exact beats I want to nail.
Get the S2.1 Pro model (free until end of July)
The quality jump comes from the S2.1 Pro model, it placed number one against every major voice tool in a blind test. Right now API access to S2.1 Pro is free for all developers until the end of July, so this is the window to build with it before it costs anything.
🎣 Open Fish Audio and clone your voice free
Tap in, clone your voice, and try one hand tagged line and one Auto Tagged line back to back. You will hear the difference immediately.
Related guides
Guide
Film Once, Post All Week: The Exact Repurpose Stack
One recording becomes clips, 30 captions, an 8-slide carousel, and a 10-tweet thread. One tool per job, the exact prompts included, and the one job you should never hand to AI.
Guide
The Best AI for Every Visual Task: One Free Tool Does 80% of the Work
Seven visual jobs, each mapped to the tool that actually wins it. The workhorse taking most of the list is free inside Gemini, and the thumbnail tiebreaker is one 12-year-old question you ask ChatGPT.
Guide
The Free, Local ElevenLabs (Clone Any Voice for $0)
OmniVoice Studio is the open-source ElevenLabs. Clone any voice from 3 seconds across 646 languages, design voices, dub videos, all on your own machine for free. Claude Code installs the whole thing for you.
System
Clone Any Viral Reel, Frame for Frame (the AI Does the Edit)
Paste a reel link. AI reverse-engineers why it went viral, hands you a frame by frame shot list, then edits your own footage into the same video: same cuts, same text behind your head, same music hits. 🤯
Get the next one first
New prompts every week.
Free. The new drops and the tools behind them, before they hit the feed.
No spam · New issues Sunday · Unsubscribe anytime
Need it custom?
Want this built for you?
Tell me the idea and I’ll build it. An app, a tool, an automation. You don’t need to be technical.
Frequently asked questions
You can clone a voice and generate speech on the free tier, and API access to the S2.1 Pro model is free for all developers until the end of July. Open Fish Audio here to start without paying.
About 10 to 30 seconds of clean audio is enough for a solid clone. Record somewhere quiet and speak naturally with a little range, the cleaner the input the more convincing the output.
No. The playground is point and click, you clone a voice and type your script right in the browser. The free S2.1 Pro API is there if you want to wire it into an app, but you never have to touch code to use the voice.
Cloning your own voice, or a voice you have explicit permission to use, is fine. Impersonating a real person to deceive people is not, and a lot of places have laws about it now. Keep it to your own voice or consenting ones and label AI audio when it matters.
ElevenLabs is the name everyone knows, but Fish Audio's S2.1 Pro model placed number one in a blind test and its inline emotion tags give you finer control over delivery. If you want a free local option too, see the free ElevenLabs alternative.