Choose a replacement AI voice tool by use: developers building live voice agents need real-time speech models, video makers need voiceover tools, and everyone else may only need a text-to-speech app. Check that the vendor is still trading, test it with your own script, and read the commercial licence before you publish anything.
LMNT's own site now says its speech generation journey has come to an end. If you'd built a product or a content workflow on it, that's a bad morning.
So if you're looking at Fish Audio alternatives, start with a vendor that's still trading, then match the tool to the job. The nine tools below are in no particular order, grouped by use: real-time voice for developers, voiceover for video and content, and everyday text-to-speech apps. Most are small independent companies. None of this is a judgement on voice quality, which you'll want to hear for yourself.
Is the vendor still trading?
Voice start-ups can disappear quickly, and when they do, your saved voices, settings and API calls go with them. Before you build on any tool here, look for recent product updates, a working support channel and clear terms about what happens to your data and audio if the service closes. Keep your own copies of scripts and finished audio files, wherever you make them.
Real-time voice for developers
Cartesia covers text-to-speech, speech-to-text and voice agents, and the vendor says its models are fast enough for live conversation from one API. Rime builds voice models made for real-time conversations in many voices and languages, pitched at enterprise teams.
Murf AI covers speech generation and conversational agents on one platform, so it spans this group and the voiceover one. If you're building a voice agent, the conversation design matters as much as the voice, and my list of the best chatbot builders covers that side.
Voiceover for video and content
Narakeet creates voiceovers and narrated videos from text or slides, so you don't have to record your own voice or sync picture and sound. Typecast is an AI voice generator with control over emotion, speed and dynamics, which helps when a flat read won't do.
WellSaid makes voiceovers with AI voices made with professional voice actors, and the vendor says it's the only AI voice generator built that way. Fliki turns text, scripts and blog posts into videos with AI voices, so the voice and the video come out of one tool.
Everyday text-to-speech apps
Voicemaker converts text to speech and lets you download MP3 or WAV files. Speechify reads PDFs, books and web pages aloud, which makes it a reading tool first rather than a production tool. For a wider view, see my guides to the best text-to-speech software and the best AI voice generators.
The nine voice tools compared
| Tool | Kind of tool | Good for |
|---|---|---|
| Cartesia | Real-time voice AI | Fast voice for live conversation |
| Fliki | Text to video with AI voices | Text to video with voices |
| Murf AI | Voice and agents platform | Speech plus voice agents |
| Narakeet | Text to speech and narrated video | Narrated videos from slides |
| Rime | Conversational voice models | Conversational voices for enterprise |
| Speechify | Text to speech reader | Reading documents aloud |
| Typecast | Expressive AI voice generator | Expressive voiceovers |
| Voicemaker | Text to speech converter | Quick downloadable audio |
| WellSaid | Studio voiceover AI | Studio-style voiceovers |
Each voice tool in a sentence
Alphabetical, and in no particular order of preference.
Cartesia
Best for: Developers building live voice agents who want text-to-speech, speech-to-text and agents through one API.
Fliki
Best for: Turning scripts, text and blog posts into finished videos with AI voices in one tool.
Murf AI
Best for: Teams that want speech generation and conversational voice agents on the same AI voice platform.
Narakeet
Best for: Making narrated videos and voiceovers from text or slides without recording your own voice.
Rime
Best for: Enterprise teams needing conversational voice models in many voices and languages for real-time conversations.
Speechify
Best for: Listening to PDFs, books and web pages read aloud in natural voices.
Typecast
Best for: Creators who want control over emotion, speed and dynamics in AI voiceovers, with an API available.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Voicemaker
Best for: Quick text-to-speech conversions downloaded as MP3 or WAV files, with a free plan stated on its site.
WellSaid
Best for: Voiceover work using AI voices made with professional voice actors.
My advice for creators and developers
Test with your own script, not the demo sentence on the homepage. Use a paragraph with names, numbers, a question and a bit of emotion, because that's where voices trip up. Listen on phone speakers as well as good headphones, since most of your audience won't be wearing studio gear.
Developers should also measure response speed inside their own app rather than the vendor's playground, and check how the service behaves when lots of requests arrive at once. Creators should export a finished file and drop it into the video editor they'll use for real, because timing and tone can feel different once there's picture on top.
Licences, consent and cloned voices
Read the commercial licence before you publish anything made with an AI voice. Check whether you can use the audio in ads, client work and products you sell, and whether the rights change if you stop subscribing.
If a tool lets you clone a voice, only clone your own or one whose owner has given you clear, written permission. Keep that consent on file, and never recreate a real person's voice without it. Rules on voice likeness and synthetic media vary by country, so check your local regulator's guidance before you publish cloned audio.
Do you make AI voice or text-to-speech software? Tool makers can ask to be considered for a featured place on this list.
Want your tool on this list? Featured spots and reviews here are paid partnerships. Here's how to partner with me.
Frequently asked questions
What's the difference between real-time voice AI and a voiceover tool?
Real-time voice AI is built to speak with very little delay inside apps and voice agents, usually through an API. Voiceover tools focus on producing finished audio files for videos, courses and content.
Can I use AI voices in commercial videos?
Often, but it depends on the vendor's licence and your plan. Read the commercial terms before publishing, especially for ads and client work.
Can I clone someone else's voice?
Only with that person's clear, written consent, and never to mislead anyone. Laws on voice likeness differ between countries, so check your local regulator's guidance.
How do I check an AI voice start-up is still trading?
Look for recent product updates, a working support channel and an active status page. Keep your own copies of scripts and audio, so you can move if a service closes.
Which tools here are built for developers?
Start with the real-time group, since those tools are built for live conversation through an API. Test response speed inside your own app.
Are these voice tools ranked?
No. They're in no particular order and grouped by use. Voice quality depends on your script and language, so test each one yourself.
