In short: The newest update brings HD voices, speaking styles and real pauses to Text to Speech (October 2026). Before that came Speechma 2.0. It rebuilds AI Dubbing (upload a video or paste a link, dub into 9 languages with natural native voices, and download the MP4), adds lyrics and sung songs in 50 languages, a Vocal Isolator that gives both vocals and a music-only track, MP3 or WAV downloads, SRT subtitles from Speech to Text, and results that stay after you refresh the page.
October 2026: HD voices, speaking styles and real pauses
Text to Speech got three new features that give you more control over how your audio sounds.
- HD voices: six new premium voices (Adam, Rowan, Marcus, Elliot, Felix and Orion) that sound extra natural. They come in English, as Multilingual voices and in about 70 other languages, with names like “Orion HD” or “Orion French HD”. This brings the total to over 1,000 voice options. Pitch, speed and speaking styles are not available for HD voices.
- Speaking styles: some voices can now speak in a style such as cheerful, sad, whispering, excited, newscast or customer service. When a voice supports styles, a Speaking Style menu appears in Voice Effects. Try Aria, Jenny, Guy, Andrew Multilingual, Sonia, Ryan, Denise, Henri, Swara or Xiaoxiao.
- Real pauses: the Short, Medium and Long pause buttons in Voice Effects now insert
[break 1s],[break 2s]or[break 3s]into your text. You can type your own too, such as[break 5s]or[break 500ms], up to 10 seconds each and up to 50 per text. Pauses work with every standard voice, including HD voices, and count toward the 2,000-character limit.
October 2026: Speechma 2.0
Speechma 2.0 (version 52) touches almost every tool. Here is what changed, tool by tool.
AI Dubbing, rebuilt
- File or link: upload an MP4, MOV or WEBM video up to 60 MB, or paste a direct video link ending in .mp4, .webm or .mov. Short clips dub fastest.
- 12 spoken languages in, 9 out: the video can be spoken in English, Hindi, Spanish, French, Portuguese, Italian, German, Russian, Japanese, Korean, Chinese or Arabic, and dubbed into English (US), English (UK), Spanish, French, Hindi, Portuguese (Brazil), Italian, Japanese or Chinese.
- Natural native voices: Automatic picks a natural native voice for the target language, or you can pick a specific voice under More options.
- Speech speed: make the dubbed speech slower or faster, from 0.8x to 1.25x.
- Video in, video out: you get the dubbed video, with the speech replaced in the new language, as an MP4 download. Your video is sent for processing only and is never stored on our servers. See the AI Dubbing page.
Songs and lyrics in 50 languages
- Lyrics follow your topic: the Lyrics tool writes original lyrics about what you describe, in 50 languages, as a short or full song.
- Songs sung in 50 languages: the Song tool sings your lyrics, or lyrics it writes from your topic, in 50 languages.
Music and Vocal Isolator
- Music on an open-source AI model: describe the music, choose 30 seconds to 8 minutes, and optionally add a reference track whose sound the music should follow.
- Vocal Isolator gives two files: the vocals and a music-only track, both ready to download. Try it on the Vocal Remover page.
Text to Speech and everyday improvements
- Results stay after refresh: your audio is kept in your own browser, so a page refresh no longer loses your work.
- Share sends the real audio file: on a phone, Share opens the share sheet so you can send the file to WhatsApp, Telegram or email.
- Open text files: load a .txt, .docx, .srt, .md, .html or .rtf file straight into the Text to Speech box.
- MP3 or WAV: a new download menu (the arrow next to each result) lets you pick the format.
- SRT subtitles from Speech to Text: turn on timestamps and download a subtitle file for your video.
- Two speakers dialogue mode: voice a conversation with two different voices in one file (details below).
Where to find each new feature
| Feature | Where it is | Learn more |
|---|---|---|
| HD voices | Text to Speech voice list (names ending in HD) | Voices |
| Speaking styles | Text to Speech › Voice Effects › Speaking Style | Make AI voiceover sound human |
| Real pauses | Text to Speech › Voice Effects › pause buttons | Help Center |
| Dub a video into another language | AI Dubbing tool | How to dub a video |
| Song in another language | Music Studio › Song Generator | AI Song Generator |
| Vocals plus music-only track | Music Studio › Vocal Isolator | Acapella Extractor |
| WAV download | Download arrow next to a result | Text to MP3 |
| SRT subtitles | Speech to Text, with timestamps on | YouTube transcription |
| Two voices in one file | Text to Speech › Two speakers | Two Speakers guide |
| Open a text file | Text to Speech text box | PDF to Audio |
October 2026: Two speakers
Text to Speech can now voice a conversation. Turn on Two speakers, start each line with A: and B: (or two names such as Host: and Guest:), pick a voice for each speaker and get one audio file with both voices, levelled to the same loudness with natural pauses. It works with standard and AI voices. Read the AI Dialogue Generator page or the Two Speakers guide.
Earlier in 2026
- Speechma in 25 languages: the site itself can now be used in 25 languages, not only English.
- Voice Cloning, Voice Design and Voice Changer: clone a voice from a clean 5 to 30 second clip (with the speaker’s permission), describe a brand-new voice in words, or re-speak a recording in another voice. All three work in 10 languages. See Voice Cloning, Voice Design and Voice Changer.
- Music Studio: music, sound effects, songs, lyrics, vocal isolation and voice covers in one place. Start with the Music Studio guide.
Suggest a feature
Many Speechma tools started as a request from a user. Tell us what you want next and vote on ideas in the Speechma Community, and get the free newsletter to hear about new features first. Found a bug? Use the contact page.
Frequently asked questions
What is the newest Text to Speech update?
HD voices (Adam, Rowan, Marcus, Elliot, Felix and Orion), speaking styles such as cheerful, sad or newscast on voices that support them, and real pauses with tags like [break 2s] that work with every standard voice.
What is new in Speechma 2.0?
Speechma 2.0 (October 2026) rebuilt AI Dubbing (video upload or link, 9 dub languages, natural native voices and an MP4 download), added lyrics and songs in 50 languages, a Vocal Isolator with vocals and music-only tracks, MP3 or WAV downloads, SRT subtitles from Speech to Text, file opening in Text to Speech and results that stay after refresh.
Are the new features free?
Yes. All tools are free with no sign-up, and commercial use of your results is included.
Does AI Dubbing store my video?
No. Your video is sent to the AI for processing only and is never stored on Speechma’s servers.
Can I download WAV instead of MP3?
Yes. Use the download arrow next to a result and choose WAV.
How do I find out about new features?
Check this page or sign up for the free newsletter. You can also suggest and vote on features in the Community.
The Speechma newsletter
Make better voiceovers, faster
- New voices and features the day they launch
- Script templates and tricks that make AI voices sound human
- Creator stories, community picks and early access
