Speechma Help Center

Text to speech and downloads

How do I add pauses to my audio?

Use pause tags. In Voice Effects, the Short, Medium and Long pause buttons insert [break 1s], [break 2s] or [break 3s] where your cursor is. You can also type your own, such as [break 5s] or [break 500ms] (up to 10 seconds each and up to 50 per text). Pause tags work with every standard voice, including HD voices, and they count toward the 2,000-character limit. Example: “First, preheat the oven. [break 2s] Then mix the batter.”

Punctuation still shapes the rhythm: a comma gives a short breath, a full stop a slightly longer one, and a new sentence adds a natural break.

Can a voice sound cheerful, sad or like a newsreader?

Some voices can. When a voice supports speaking styles, a Speaking Style menu appears in Voice Effects with options such as cheerful, sad, whispering, excited, newscast or customer service. Voices with styles include Aria, Jenny, Guy, Andrew Multilingual, Sonia, Ryan, Denise, Henri, Swara and Xiaoxiao. Other voices keep their normal delivery.

What are HD voices?

HD voices are premium voices that sound extra natural: Adam, Rowan, Marcus, Elliot, Felix and Orion. They are available in English, as Multilingual voices and in about 70 other languages, with names like “Orion HD” or “Orion French HD”. Pitch, speed and speaking styles are not available for HD voices, but pause tags work.

How do I download all my clips at once?

After generating several clips, use Download all to save them together as a single ZIP file. This is handy when you narrate a presentation slide by slide or a course lesson by lesson.

Where is my history saved?

Your history is kept privately in your own browser on your device. It is not tied to an account. Clearing your browser data, switching browsers or using a private window will remove it, so download any audio you want to keep.

What are the character voice styles?

Besides the 1,700+ standard voices, Speechma has character voices in styles such as narrator, trailer, documentary, news and storyteller. Some of them read up to 1,000 characters per generation. Try a trailer style for a movie trailer voice.

Why does a number or name sound wrong?

Write it the way it should be spoken. For example, “twenty twenty-six” instead of “2026”, “Doctor” instead of “Dr.”, or a phonetic spelling for unusual names.

Voice cloning, voice design and voice changer

What kind of sample do I need for voice cloning?

A clear recording of one person speaking for 5–30 seconds. You can upload any common audio format or paste a link. Scripts can be up to 1,000 characters, in English, Spanish, French, German, Italian, Portuguese, Russian, Chinese, Japanese or Korean, with speed from 0.8× to 1.25×. See Voice Cloning.

Can I clone a celebrity or someone else’s voice?

No. Only clone your own voice or a voice you have clear permission to use. Impersonating real people breaks our AI Voice & Cloning Policy. If you want a distinctive character, use Voice Design to create an original voice instead.

How does Voice Design work?

Describe a voice in words, such as age, gender, tone, accent and mood (for example, “a friendly young woman with a light Texas accent, upbeat and warm”), add your script, and Speechma’s AI creates a brand-new voice. It supports the same 10 languages.

What does the voice changer do?

The Voice Changer takes your speech recording (up to 5 minutes) and a target voice sample (5–30 seconds), then re-speaks the same words in the target voice. It works in the same 10 languages.

AI dubbing and speech to text

Which videos can I dub?

Upload an MP4, MOV or WEBM file up to 60 MB (short clips dub fastest), or paste a direct link ending in .mp4, .webm or .mov. Pick the spoken language (12 supported: English, Hindi, Spanish, French, Portuguese, Italian, German, Russian, Japanese, Korean, Chinese or Arabic), then dub into one of 9 languages: English (US), English (UK), Spanish, French, Hindi, Portuguese (Brazil), Italian, Japanese or Chinese. Leave the voice on Automatic for a natural native voice, or pick one under More options and set the speed from 0.8x to 1.25x. You get the dubbed video as an MP4. Your video is sent for processing only and is never stored on our servers. See AI Dubbing.

How long can my audio be for speech to text?

Up to about 1 hour, in formats such as MP3, WAV, M4A, OGG and FLAC, in many languages. Choose plain text or timestamps per sentence.

Can I get timestamps for subtitles?

Yes. Choose plain text or timestamps per sentence. You can copy the result or download it as text, and with timestamps on you can download an SRT subtitle file ready for YouTube or social video.

Music Studio

Which tools are in Music Studio?

Six: AI Music Generator, Sound Effects, Song Generator, Lyrics Generator, Vocal Remover and AI Voice Cover. All are free with no account, and each generation needs a 5-digit security code.

How long can AI music tracks be, and in what format?

Instrumental tracks from 30 seconds to 8 minutes. Download MP3, or WAV with the arrow next to Download. You can add an optional reference track, and the music borrows its sound and feel. See the AI Music Generator.

Can I make a song with my own lyrics?

Yes. The Song Generator uses your lyrics, or writes lyrics from a topic, plus a style description. Songs can be 30 seconds to 8 minutes, with an instrumental option, sung in 50 languages. Need words first? Try the Lyrics Generator (50 languages, short or full song).

How do sound effects and the vocal remover work?

For sound effects, describe a sound and get a 3–15 second clip. The Vocal Remover splits a song or video up to about 9 minutes into a vocals track and a music-only track, and you can download both.

Why does AI Voice Cover take longer?

Re-singing a whole song in a new voice style (female pop, male pop, soft female, deep male, powerful female, rock male, rap male or choir) is heavy work and takes a few minutes. Keep the tab open until it finishes. Only upload songs you have the rights to use.

Devices, app and browsers

Is there a Speechma app?

Yes, an Android app on Google Play. Search for “Speechma”.

Does Speechma work on iPhone and iPad?

Yes, through the website. Speechma works in any modern browser on desktop and mobile, including Safari, Chrome, Edge and Firefox.

Troubleshooting and errors

What does a reference code like SM-XXXXXXX mean?

When something goes wrong, Speechma shows an error with a reference code. Copy it exactly and email it to contact@speechma.com with the tool name, your browser and what you tried. It lets us locate the problem quickly. See our contact page for tips.

My security code is not accepted. What should I do?

Type all 5 digits exactly as shown. If it still fails, load a new code and try again. Browser extensions that block scripts can sometimes interfere.

My generation is slow or did not finish. What should I do?

At busy times a generation can take a little longer. Keep the tab open, and if it fails, load a new security code and try again. If you want no captcha, no ads and no waiting, SpeechmaPRO.com is our premium option.

My upload was rejected.

Check the file requirements for the tool: 5–30 seconds for cloning samples, up to 5 minutes for voice changer recordings, 60 MB for dubbing videos (MP4, MOV or WEBM), about 1 hour for transcription and about 9 minutes for the vocal remover. Convert unusual formats to MP3 or MP4 and try again.