OnlyVoice

First, type or paste your text into the text-to-speech editor and mark it up where needed. You can add a pause of a set length between phrases, insert a role, replace the pronunciation of a word and adjust reading settings, with undo and redo as you go.

Second, choose the voice. You can browse the voice library and filter it by language, use a voice you cloned by uploading or recording up to 20 seconds of audio (MP3, WAV, MP4 and other formats are accepted), or create a new voice in Voice Design by describing it in a prompt and previewing it on sample text.

Third, shape the delivery with the emotion, speaking speed, tone and volume controls, then generate, listen and download the file in the format you need. When you are signed in, every result is kept in your generation history; audio generated while logged out is shown only temporarily and is not saved.

• •Text to Speech Editor - Pauses with set lengths, roles, pronunciation replacement and reading settings give you control over how a script is read. The character limit per generation depends on the plan, from 150 on Free to 1,000, 5,000 and 10,000 on the paid plans.
• •Voice Cloning - Upload or record a 10 to 20 second sample and the AI learns the voice in a few seconds. Optional audio preprocessing reduces background noise, enhances speech and normalizes volume, and can be switched off for clean recordings. Free includes one clone; paid plans allow 5, 10 or 30 voice clone models.
• •Voice Design - Write a description of the voice you want, up to 500 characters, and preview it on your own text. No recording is needed, which helps when you want a character voice rather than your own.
• •Voice Library - More than 5,000 voices in 11+ languages, labelled as natural, designed or cloned, with a language filter so you can find a narrator for multilingual content.
• •Delivery Controls - Emotion, speaking speed from slower to faster, tone from deeper to brighter and volume from quiet to strong let you fit the voice to the context instead of accepting a default read.
• •Common Export Formats - Download results as MP3, WAV or AAC. Commercial use is listed on the paid plans, starting with Spark.

YouTubers and video creators can produce voiceovers in several languages from one script, and keep the same voice across a whole channel by cloning their own or designing one. Course creators and educators can narrate lessons without re-recording when the text changes, using pauses and pronunciation fixes to keep technical terms right.

Podcasters and marketers can add intros, ads and narration, and product teams can give an app or demo a consistent spoken voice. Because the paid plans are short passes with no automatic renewal, the tool also suits people with a one-off project who want a month of higher limits without managing a subscription. The terms require that you have permission for any voice sample you upload, so cloning is meant for your own voice or a voice you are authorized to use.