Acoust AI
Acoust AI is a browser-based text-to-speech studio that turns scripts into lifelike voiceovers, clones a voice from a 10-second sample and bundles a video editor, so the audio and the finished video come out of one tool.
What is Acoust AI?
Acoust AI is a browser-based platform for turning written text into spoken audio, and then into finished video. It is published by Bytebot LLC, which trades as Acoust AI, and it runs as a web application with nothing to install.
The core of the product is text-to-speech. You type, paste or import a script and the platform reads it back in a synthetic voice built on generative language models; cloning and synthesis are credited to Google's Gemini models. Delivery is adjustable rather than fixed: pauses, laughter, breathing, pitch, speed, emphasis, whispers and intensity can all be dialled in, custom pronunciations can be stored, and a voice can even be conjured from a short prompt describing the tone you are after, from a calm narrator to an energetic host.
Voice cloning is the second pillar. A consent recording plus a clean ten-second sample produces a digital copy of a voice, which then works across languages, so one narrator can carry a whole catalogue into several markets.
Around these sit a video editor and a clip generator, both still labelled beta: the editor assembles images, clips and audio on a timeline with a stock media library, while AI Clips turns long recordings into short-form pieces with automatic subtitles. Translation, audio and YouTube transcription, SRT subtitle export, an AI writer and a slides-to-video converter round out the set.
Language coverage is advertised loosely, with pages variously claiming 30+, 40+ or 60+ languages and between 100 and 250 voices. The grid actually published lists forty language and accent combinations covering thirty-three distinct languages, with regional variants for English, Arabic, French, Portuguese and Spanish.
The audience is broad: content creators, learning and development teams, teachers, marketers, real estate agents, online sellers and support teams building phone menus. Twenty-one use cases are documented, from audiobooks and podcasts to Discord bots and video games. Teams get pooled credits, per-seat pricing and single sign-on on the Enterprise tier. One thing is conspicuously missing: the site publishes no privacy policy, and nothing beyond the legal name identifies the company behind it.
What it does
- Turn any script into a lifelike AI voiceover, choosing from the published voice catalogue
- Clone your own voice from a 10-second sample and reuse it across languages
- Create a custom voice by describing the tone, pacing and mood you want in plain text
- Fine-tune delivery with pauses, laughter, breathing, pitch, speed, emphasis and whispers
- Cut a long video into short social clips with styled automatic subtitles
- Assemble a finished video on a timeline with stock footage, images and your generated audio
- Transcribe audio or a YouTube video, translate the text, and export SRT subtitles
When to use Acoust AI / When not to
A quick filter to help you decide if Acoust AI is the right fit.
When to use Acoust AI
- YouTube, TikTok and Reels creators who need broadcast-ready voiceovers without a microphone or a studio
- Learning and development teams producing e-learning modules, onboarding and compliance training that must be updated without re-recording
- Teachers, lecturers and course designers turning written lessons into narrated audio for their students
- Real estate agents and online sellers narrating property tours, listings and product videos at volume
- Small marketing teams and solo founders localising ads, explainer videos and IVR prompts across dozens of languages on a tight budget
When not to use Acoust AI
- Organisations processing personal data under the GDPR: the site publishes no privacy policy and never mentions the regulation
- Regulated sectors that need a Data Processing Agreement, a named subprocessor list or a documented hosting jurisdiction, none of which exists here
- Developers looking for a public API: the terms mention API keys, but no documentation is published anywhere
- Mobile-first users, since there is no iOS or Android application and everything happens in the browser
- Anyone hoping to publish commercially from the free plan, which blocks audio exports and forbids commercial use
How to use Acoust AI
A typical end-to-end flow, from setup to results.
- Create an account on the Acoust web app; the free plan asks for no credit card
- Open a new project from the dashboard and enter the editor
- Type, paste or import your script, from a YouTube link, a web page, a text file, a PDF or a .docx document
- Pick a voice from the catalogue, or describe the voice you want in a short text prompt
- Adjust the delivery with pauses, emphasis, speed and custom pronunciations; plain text is enough and SSML is optional
- To use your own voice, record the consent statement and a clean ten-second sample in a quiet room, then wait for the clone to appear
- Generate the audio and listen back before committing to an export
- Optionally build a video: add clips and images, align them with the audio on the timeline and draw on the stock library
- Export the audio as MP3 or share a hosted link; exporting requires a paid plan
- Consult the help centre for guides on projects, voices, billing and SSO, or email support on weekdays
Pros & Cons
Pros
- Text-to-speech, voice cloning and video editing bundled into a single subscription
- Voice cloning from a ten-second sample, free to try before paying
- A permanent free plan that needs no credit card
- Aggressive pricing: 9 US dollars a month buys 180 minutes of AI voice, 29 dollars buys 600
- A commercial licence is included from the first paid tier upwards
- Forty published language and accent combinations, including several regional variants of the same language
- No lock-in: plans can be switched or cancelled at any time, and 5-dollar top-up packs absorb a busy month
Cons
- No privacy policy is published anywhere on the site, which is the single most serious gap
- No GDPR mention, no Data Processing Agreement, no named subprocessors and no EU representative
- Nothing is disclosed about where data is hosted, or about whether customer content trains the models
- Marketing figures contradict each other from page to page: 100+, 200+ or 250 voices, 30+, 40+ or 60+ languages
- The free plan blocks audio exports and forbids commercial use, so nothing produced on it can be published
- The video editor and AI Clips are still in beta, and no public API is documented despite the terms mentioning API keys
- One seat only on both Pro and Premium, no mobile application, and email-only support on weekdays
Pricing & Plans
Acoust AI operates on a freemium basis. A permanent free plan, named Personal, is available without a credit card, but it is capped at 10,000 credits, roughly ten minutes of AI voice, and it blocks audio exports as well as any commercial use. The lowest paid entry point is the Pro plan at USD 9.00 per month on monthly billing, falling to USD 7.00 per month when billed annually at USD 84.00 per year.
- free
- 10
- 000 credits
- around 10 minutes of AI voice
- AI Writer
- cloud storage and premium GenAI voices
- but no audio export and no commercial use
- USD 9 per month
- or USD 7 per month billed annually at USD 84 per year
- with 180
- 000 credits a month covering up to 180 minutes of AI voice or 90 minutes of cloning
- SRT subtitles
- a commercial licence
- 1 seat
- 50 projects and 10 GB of storage
- USD 29 per month
- or USD 22 per month billed annually at USD 264 per year
- with 600
- 000 credits a month covering up to 600 minutes of AI voice or 300 minutes of cloning
- AI Clips
- 60 minutes of transcription
- AI translation
- 1 seat
- custom pricing
- with team accounts
- pooled credits
- per-seat pricing
- SSO
- custom quotas and dedicated support
- USD 5 for 25 additional AI prompts
- USD 5 for 50
- 000 additional characters
Data, GDPR & hosting
A consolidated view of how Acoust AI handles your data.
GDPR overview
There is nothing to report, and that absence is the finding. Acoust AI publishes no privacy policy at all: /privacy, /privacy-policy and /legal all return 404, and none of the 176 URLs in the sitemap points to one. The word GDPR appears nowhere across the fourteen pages collected. No Data Processing Agreement is offered, no subprocessor list is published, no Article 27 EU representative is designated, and no data protection officer or dedicated privacy mailbox exists, hi@acoust.io being the only address on the domain. The terms place the contract under California law, with binding AAA arbitration in Los Angeles County and a waiver of jury trial and class actions. The only privacy-adjacent claim is a marketing line saying the platform is designed to meet all security and compliance requirements, which is substantiated nowhere.
Who owns the data?
The terms are unusually direct on ownership. Section 3 states that you own the applications and the data you submit, while Bytebot LLC keeps every right in the platform itself and its content. Section 5 adds that you retain the rights to your Input, own the Output, and that Acoust AI assigns to you any rights it might hold in that Output. In return you grant the company a limited licence to process and display both, purely so the service can run. A cloned voice stays under your control and can be deleted at any time. The one caveat is commercial: only paid plans carry a commercial licence for the audio you generate.
Reuse rights
You can reuse what the platform produces without asking permission, within two boundaries. The first is commercial: the free Personal plan is marked "Not for Commercial usage" and blocks audio exports altogether, so a commercial licence only arrives with a paid plan, and it then covers videos, advertising, podcasts and client work. The second is ethical and contractual: the Responsible Voice Use Guidelines allow you to upload or clone a voice only where you hold valid rights and documented consent, forbid impersonation of anyone living or dead, and require you to disclose that audio is AI-generated whenever it could be mistaken for a real person. On its own side, Acoust AI limits itself to processing and displaying your content to deliver the service, and states that personal data is not shared for marketing or profiling purposes.
Data retention & training
Hosting summary
Acoust AI discloses nothing about where data is hosted. No country and no region is named on any of the pages collected, and there is no privacy policy or trust page where such information would normally sit. The terms say only that audio generation may be handed to trusted third-party providers acting as data processors on behalf of Bytebot LLC, contractually bound to secure the data, use it solely to fulfil the request and delete temporary copies once processing is complete; none of those providers is named. The voice cloning page credits Google's Gemini models for cloning and synthesis, which implies at least one external processor but says nothing about its location. The marketing site itself is served behind Cloudflare, whose anycast address resolves to a node in the United States, an infrastructure detail about the website rather than a statement about where projects and voice samples are stored. The contract is governed by California law. Anyone with a residency requirement should treat the hosting question as unanswered and put it to the vendor in writing.
Where Acoust AI works
Country-level availability.
Not available in
Things to keep in mind
Risks and trade-offs to weigh before adopting Acoust AI.
- With no privacy policy at all, you cannot know what is collected, where it travels or how long it is kept
- Audio generation is delegated to third-party providers the terms never name, so you are trusting a chain you cannot inspect
- The site takes no explicit position on whether customer content trains its models, and offers no opt-out mechanism
- Voice cloning demands verifiable consent: cloning a colleague, a client or a public figure without written permission exposes you legally, whatever the tool lets you do technically
- Synthetic narration that is not disclosed can mislead an audience; the guidelines require disclosure, but nothing in the product enforces it
- Liability is capped at USD 100 or twelve months of fees, and disputes go to arbitration in Los Angeles with no jury and no class action
- Leaning on generated voices can quietly erode a team's own recording, scripting and editing skills, and the free plan's export ban means a project can be finished before you discover you cannot publish it
Setup & Integrations
Technical difficulty
Very low. Acoust AI runs entirely in the browser: nothing to install, no key to configure, no integration to wire up. Signing up takes an email address and the free plan asks for no card. The video editor is presented as requiring no prior experience, and SSML is optional since plain text is enough. The only real preparation concerns voice cloning, which needs a consent recording and a clean ten-second sample captured in a quiet room. Only the Enterprise tier adds a genuine configuration step, with single sign-on to set up alongside team seats and quotas.
Deployment
Integrations
Supported languages
Behind Acoust AI
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Acoust AI.
Frequently asked questions
What is Acoust AI?
Is there a free plan, and what does it leave out?
How much does the first paid plan cost?
How many voices and languages are really available?
How does voice cloning work?
Can I download the audio and use it commercially?
Is there an API or a mobile app?
What does Acoust AI say about privacy and the GDPR?
Should you pick Acoust AI?
Acoust AI packs an unusual amount into one low-priced subscription. Nine dollars a month buys expressive text-to-speech, voice cloning from a ten-second sample, translation, transcription, subtitles and a video editor, a combination that normally means paying for two or three separate tools. The permanent free plan lets you hear the voices before spending anything, the commercial licence arrives with the very first paid tier, and 5-dollar top-up packs mean a busy month does not force an upgrade. For a solo creator, a small marketing team or a training department that has to refresh forty modules without booking a studio, the value on offer is real.
The reservations are just as clear, and they sit almost entirely on the governance side. There is no privacy policy, none at all, on any URL, which means no stated retention period, no hosting jurisdiction, no subprocessor list, no DPA and no position on whether your scripts and voice samples feed model training. Audio generation is handed to unnamed third parties. The GDPR is never mentioned. The company publishes no postal address and names no officer. Two of the four modules are still in beta, the advertised voice and language counts contradict each other from page to page, and a leftover FAQ still answers that team accounts do not exist while the rest of the site sells them.
The fair reading is a capable production tool with immature paperwork. If the material is your own and your stakes are commercial rather than regulatory, Acoust AI is worth a serious trial. If you handle personal data or expect to answer a procurement questionnaire about where recordings are stored, put those questions to the vendor in writing first: the website will not answer them for you.
- Choosing a selection results in a full page refresh.
- Opens in a new window.