Hey Sumo-lings 👋

Big month at Vocallab. Here's what shipped:

• Video dubbing (new) — Upload a video, pick a language, and get it back re-voiced with the timing preserved, plus an SRT track. It's in beta and live in the Studio now.
• Speech to Text up to 500MB — The old 18MB cap meant long recordings had to be split by hand. Full podcasts, interviews and lecture recordings now go up in one piece.
• Voice cloning and voice design in the API and MCP — Anything you can do in the Studio, you can now do from your own code or straight from an AI assistant. We also widened language and accent support across the board.
• SRT caption export in the API and MCP — Captions are no longer a Studio-only feature.
• Audiobooks got smarter — The engine that reads your manuscript and works out who's speaking is meaningfully better at it, we fixed a batch of bugs you reported, and every voice in the cast picker now has a preview you can play before you commit.
• Faster, cleaner UI — A lot of unglamorous work on load times and rough edges throughout the app.

💙 Thank you for the bug reports and feature requests. Most of this list came from you, so keep them coming. 💙

— The VocalLab team

Hey Sumo-lings 👋

Two big things landed this month — and one of them came straight from your requests.

🎧 Turn any ebook into an audiobook (brand new)

Upload an EPUB, PDF, DOCX, or TXT and VocalLab will:

- Automatically detect the narrator and every speaking character

- Let you cast a voice to each one — or auto-cast the whole book in one click

- Generate the full book with emotion and natural, expressive delivery

- Export the finished audiobook as a single file

You don't have to sit and wait: close the tab and we'll email you the moment it's ready to cast, and again when the audio is finished.

🎙️ Improved voice cloning

We reworked the cloning pipeline for more accurate, natural-sounding clones from the same short sample.

✅ You asked, we shipped

A batch of audiobook-editor upgrades, all from your feedback:

- Download individual chapters

- Delete chapters you don't need

- Reorder and move chapters around

- …plus a bunch of quality-of-life fixes throughout the editor

Everything here is included in your lifetime deal — no extra cost and no new tiers.

Keep the requests coming in the comments — a lot of what you see above started as a single message from a Sumo-ling. More on the way 🚀

— The VocalLab team

Hey Sumo-lings 👋

First — thank you. The questions and honest feedback in this thread have directly shaped what we've built over the last couple of weeks. Here's what's new:

🎙️ Speech to Text (brand new)

Upload or record audio and get an accurate transcript — 25+ languages, auto-detected. Copy it, download as .txt, or send it straight into Text to Speech. And it's cheap: about 15 points per minute of audio — roughly 4× less than generating speech. Live in the Studio now.

✨ Voice Design just got easier

Not sure how to describe the voice you want? There's a new "Improve" button — type a rough idea ("calm British woman for sleep stories") and it rewrites it into a detailed prompt the model follows well. Great voices, no guesswork.

📚 Audiobook editor — paste and go

Paste your whole script and it auto-splits into editable blocks. No more pasting line by line.

🎛️ The best model, front and center

The VocalLab Studio model (our most natural, steerable engine) is now easy to find and selected by default where your plan includes it — no more digging through Advanced settings.

🌍 Cloning in 90+ more languages

Pick "Other" when cloning and it auto-detects the language from your sample.

👥 Team seats

Workspace seats are now self-serve on Pro and above — each seat adds minutes to a shared team pool.

🐛 And fixes you reported — the API voice-listing bug, MCP download blocks, and sign-in/email delivery for Yahoo/AOL — all fixed, usually within hours.

We read every comment here. Keep them coming — a lot of the above started as a single question in this thread.

Thank you for backing us 🙏

— Petr & the VocalLab team

A few updates from us 👋

Hi everyone! Quick update on what we've shipped and what's coming.

🇩🇪 "Where are the German voices?"

This one hurt a little — a few of you couldn't find our non-English voices, and some even refunded because of it. So let me say it clearly: we DO have native German voices (and native voices in many other languages too). They were there all along, just hard to find. 😅

👉 Go to https://app.vocallab.ai/library and sort/filter by language — you'll see every voice we have in that language. Please give them a try before deciding!

🔧 API & MCP fixes

The list voices API endpoint now returns all your voices — including the default ones, not just your own. ✅

We made installing MCP into Claude much easier. 🤖

🚀 Coming Monday (27 Jul 2026)

Big audiobook improvements 📚 (still in beta, but a real step up)

Speech-to-Text (STT) 🎙️

🙏 Thank you

Honestly — the questions you've been asking are what pushed us to ship all of this. You found the gaps faster than we did. Special thanks to this community, you're the best. 💙

We read every request and suggestion, and we're doing our best to bring you more joy and a better experience in VocalLab. Keep them coming!

— Petr, founder of VocalLab

Hey Sumo-lings,

Two updates to voice cloning this week 👇

1) New "Other" language option

Until now, cloning was limited to our 15 core languages. We just added an "Other" option to the language picker — pick it and the engine auto-detects the language straight from your audio sample.

This unlocks cloning in 90+ additional experimental languages (200+ when you count regional locales)

Afrikaans, Albanian, Amharic, Armenian, Assamese, Azerbaijani, Basque, Belarusian, Bengali, Bulgarian, Burmese, Cantonese, Catalan, Cebuano, Croatian, Czech, Danish, Eastern Yiddish, Egyptian Arabic, Estonian, Filipino, Finnish, Galician, Garhwali, Georgian, Greek, Gujarati, Gulf Arabic, Haitian Creole, Hijazi Arabic, Hungarian, Icelandic, Indonesian, Javanese, Kannada, Kazakh, Konkani, Lao, Latvian, Libyan Arabic, Lithuanian, Luxembourgish, Macedonian, Maithili, Malagasy, Malay, Malayalam, Marathi, Mongolian, Najdi Arabic, Nepali, Northern Uzbek, Norwegian Bokmål, Norwegian Nynorsk, Odia, Omani Arabic, Pashto, Persian, Piedmontese, Punjabi, Romanian, Serbian, Sindhi, Sinhala, Slovak, Slovenian, Swahili, Swedish, Tamil, Telugu, Thai, Tunisian Arabic, Turkish, Ukrainian, Urdu, Uzbek, Vietnamese, Võro, Welsh.

One honest note: these experimental languages are a research preview. Quality is already good, but it can be a little less consistent than our 15 core languages. Give it a try and let us know how it sounds for yours.

2) Dutch fix

Dutch was supported by the model but wasn't showing up as a selectable option in the Studio app — a gap on our side. It's now live in the language list alongside the other core languages. If you clone in Dutch, you're all set.

As always, keep the feedback coming — a lot of what we ship comes straight from your requests.

— The VocalLab team