Notta
Tool · this deep dive is built from the official pages of the platform itself and of the tools that support it.
Overview
Notta turns speech into text. Composition from the features sections and comparison table of the pricing page.
Meetings are recorded and transcribed two ways: with a bot that joins the call, and without one. Uploads take audio and video files and YouTube links; screen recording is included. Transcription runs live or from a finished recording, with bilingual and on-device transcription (nothing sent to a server) named separately. Translation comes live, bilingual, or applied to a finished transcript. Plus AI summaries and questions about meetings.
Recognition languages are where the site falls short: no exact list, only a round number. From the audio-to-text page: "Notta can convert audio recording to text in 58 languages, including English, Spanish, German, French, Portuguese, Japanese, and many more" — six examples out of fifty-eight. Of our catalogue's six languages only English and Spanish are named; Russian, Chinese, Hindi and Arabic are not. No list page exists on the site or in the sitemap: the number is known, the names are missing.
Do not confuse it with the interface, which switches to 19 languages including Arabic, Farsi, Hindi and Ukrainian, but NOT Russian.
Which platforms it works with
Integrations from the comparison table. Calls: Zoom, Microsoft Teams, Google Meet, Webex. Workspaces: Slack, Notion, Zapier. CRM: Salesforce, HubSpot, Pipedrive, Zoho CRM, Zendesk Sell, Salesflare, Freshsales. Calendars: Google Calendar, Outlook Calendar. Storage: Google Drive, Microsoft SharePoint, Dropbox, Microsoft OneDrive, Box.
Apps: web, iOS, Android, desktop for Mac and Windows, a Chrome extension.
Export is declared in two lines — "Export transcripts" and "Download transcripts or recordings" — with no format list, hence the partial mark. SRT and VTT are not declared anywhere on the pages read. Direct export to social networks does not exist: this is built around meetings, not publishing.
Pricing
From the pricing page, US dollars. A caveat: monthly prices for Pro and Business are not served on the static page — the Monthly/Annual switch runs on script, and only annual values sit in the markup. Below, what is actually published.
Pro — 8.17 USD a month annually, 97.99 USD a year, with a "40% OFF" note. Business — 16.67 USD a month annually, 199.99 USD a year. Enterprise — on request, from 51 seats.
Then paid add-ons, bought separately from the plan. Monolingual Translation — 6 USD a month annually (72 a year) or 10 monthly. Bilingual Transcription & Translation — 9 USD a month annually (108 a year) or 15 monthly. Notta Brain Pro — 93.59 USD a year for 8000 AI credits a month. A seven-day Business trial exists.
Entry minimum: 8.17 USD a month on annual Pro, 97.99 USD in one payment. The monthly figure is not shown.
What is free
Free — 0 USD, marked "No credit card required": one seat, 120 transcription minutes a month, 50 file uploads, 10 AI summaries, 1000 Notta Brain credits, speaker identification.
Then the limit found late: one conversation on Free is capped at 3 minutes. An hour-long meeting does not fit, whatever remains in the month.
Second: translation and export of a transcript do not exist on Free, both rows being marked from Pro up. Precisely — translation is unavailable in principle, not limited by volume, and the transcript cannot leave as a file either.
The seven-day Business trial is a paid plan tasted, not a free one.
Restrictions
The quota is marked "Transcription quota (no carryover)": it does not roll into the next month.
Monthly minutes: Free 120, Pro 1800, Business unlimited, Enterprise by contract. Cap per recording: Free 3 minutes; Pro, Business, Enterprise 5 hours. Uploads a month: 50, 100, 200, unlimited. AI summaries: 10, 100, 200, unlimited.
Translation counts in runs, not minutes, as the page puts it: up to 100 a month per account with the Monolingual add-on, up to 200 with Bilingual Transcription & Translation, and without an add-on up to 2 and up to 4 free runs a month.
Mind the gap between Business's unlimited quota and the 5-hour cap on one recording: an unlimited month does not lift the per-file limit.
Rights to the output
Rights to what you upload stay with you, said outright in the terms: "Customer retains all ownership rights in Customer Data".
The licence to the tool is narrow — "to access, process, copy, export, and display Customer Data, only as reasonably necessary (a) to provide, maintain and update the Services", then a closed list. "Only as reasonably necessary" ties it to delivering the service; a right to train is not there.
What the terms lacked: no clause on rights to the transcript and translation, nothing on commercial use of the result. The word "transcript" never appears in the rights clause — that it is covered follows from the definition of Customer Data. Hence the partial mark.
No requirement to label a transcript as machine-made: no watermarks, no metadata, no C2PA in the terms or the prices.
Do they train on your data
No direct answer — marked absent. Neither the terms nor the security page addresses training on uploaded recordings.
The only mention of training in the privacy policy is narrow, covering Google Workspace data alone: "Google Workspace APIs are not used to develop, improve, or train generalized/non-personalized AI and/or ML models".
It cannot be stretched over everything uploaded: it speaks of data obtained through Google Workspace interfaces and nothing else. A general denial of training on user audio does not appear.
Nor does an opt-out, in terms, policy or security page. The right to demand deletion exists, but deletion and an opt-out are different things.
How you earn with it
On commercial use the terms are silent: with no clause on rights to the transcript and translation, there is neither permission nor ban. Certain only is that rights to uploads (Customer Data) stay with the customer, and the tool's licence is confined to delivering the service.
Transcribing other people's recordings for hire carries one more weight: responsibility for the legality of the recording sits on the user outright.
What it will not do
Take responsibility for recording other people it will not. The terms: "It may be an offense in your jurisdiction to record other individuals without their prior written consent. It is your responsibility to follow applicable foreign, federal and local laws when using such features, including informing the participants in your meetings that the discussion is being recorded". For a tool that walks a bot into somebody else's call, that is the central clause.
Accept prohibited material it will not. Barred are uploads that are defamatory, obscene, offensive, hateful, pornographic, or that promote violence or discrimination by race, sex, religion, nationality, disability, orientation or age; that infringe a patent, trademark, trade secret or copyright; that disclose other people's confidential or private information; that encourage unlawful activity; or that carry malicious code.
Transcribe an hour-long meeting on Free it will not: the cap per conversation is 3 minutes. Translate a transcript free it will not: translation is absent from Free, starts at Pro, and past the 2 and 4 free runs a month needs a paid add-on. Hand you the transcript as a file on Free it will not: export is marked from Pro up. Carry unspent minutes forward it will not: the quota says "no carryover". Publish to social networks it cannot — no such export.
Verified data
The checked data this deep dive rests on.
About the platform
speech-to-text transcription: recording and transcribing online meetings with or without a bot, uploading audio and video files and YouTube links, screen recording, real-time transcription, bilingual transcription, on-device transcription without sending anything to a server, real-time translation and bilingual translation, translation of a finished transcript, AI summaries and questions about meetings
the composition is taken from the Features sections and the plan comparison table on the pricing page
source, checked 2026-08-07
THERE IS NO EXACT LIST ON THE SITE — only a round number. Verbatim from the audio-to-text page: «Notta can convert audio recording to text in 58 languages, including English, Spanish, German, French, Portuguese, Japanese, and many more» — that is, six examples out of 58 are named and the full list is not published. Of the six languages of the catalogue only Spanish and English are named outright; about Russian, Chinese, Hindi and Arabic the page is silent. The site interface switches to 19 languages (among them Arabic, Farsi, Hindi and Ukrainian, but NOT Russian) — these are interface languages, not recognition languages
no list of recognition languages could be found either on the product pages or in the sitemap; the interface languages were deliberately not mixed with the recognition languages
source, checked 2026-08-07
Platforms
integrations per the comparison table: Zoom, Microsoft Teams, Google Meet, Webex, Slack, Notion, Zapier, CRM (Salesforce, HubSpot, Pipedrive, Zoho CRM, Zendesk Sell, Salesflare, Freshsales), Google Calendar, Outlook Calendar, Google Drive, Microsoft SharePoint, Dropbox, Microsoft OneDrive, Box. Applications: web, mobile iOS and Android, desktop for Mac and Windows, a Chrome extension. Export of the transcript is claimed by the rows «Export transcripts» and «Download transcripts or recordings» with no list of formats; specific subtitle formats are not named on the pricing page
SILENCE OF THE SOURCE as to subtitle formats: SRT/VTT are not claimed on the pages read. There is no direct export to social networks — the service is built around meetings, not publishing
source, checked 2026-08-07
Pricing
yes, the Free plan — 0 USD, «No credit card required» (verbatim): 1 seat, 120 minutes of transcription a month, no more than 3 minutes per conversation, 50 file uploads a month, 10 AI summaries a month, 1000 Notta Brain credits a month, speaker identification. Translation of the transcript and export of the transcript are NOT on Free — both rows in the comparison table start from the Pro plan
important where translation is concerned: on the free plan translation is unavailable in principle rather than limited in volume
source, checked 2026-08-07
Pro — 8.17 USD a month paid yearly, 97.99 USD a year in total (marked «40% OFF» on the page); Business — 16.67 USD a month paid yearly, 199.99 USD a year in total; Enterprise — price on request, from 51 seats. Separate paid add-ons: Monolingual Translation — 6 USD a month paid yearly (72 USD a year) or 10 USD paid monthly; Bilingual Transcription & Translation — 9 USD a month paid yearly (108 USD a year) or 15 USD paid monthly; Notta Brain Pro — 93.59 USD a year for 8000 AI credits a month. There is a 7-day trial of Business
the monthly prices of the Pro and Business plans themselves are not delivered by the static page — the Monthly/Annual switch works by script; only the yearly values 8.17 and 16.67 are present in the HTML, so the monthly ones are not recorded
source, checked 2026-08-07
8.17 USD a month on the yearly billing of the Pro plan (97.99 USD a year); the monthly price is not displayed on the page that was read
source, checked 2026-08-07
Limits and restrictions
the transcription quota is marked «Transcription quota (no carryover)» (verbatim — it does not carry over): Free 120 minutes a month, Pro 1800 minutes a month, Business unlimited, Enterprise by agreement. Maximum per recording: Free 3 minutes, Pro/Business/Enterprise 5 hours. File uploads a month: Free 50, Pro 100, Business 200, Enterprise unlimited. AI summaries a month: 10 / 100 / 200 / unlimited. Translation add-ons: Monolingual Translation up to 100 times per account a month, Bilingual Transcription & Translation up to 200 times; without the add-on, up to 2 and up to 4 free runs a month respectively
the translation limits are counted in «times» rather than minutes — so says the pricing page
source, checked 2026-08-07
Restrictions
responsibility for recording someone else's speech is placed on the user outright. Verbatim: «It may be an offense in your jurisdiction to record other individuals without their prior written consent. It is your responsibility to follow applicable foreign, federal and local laws when using such features, including informing the participants in your meetings that the discussion is being recorded». It is forbidden to upload material that is defamatory, obscene, abusive, hateful or pornographic, that encourages violence or discrimination on grounds of race, sex, religion, nationality, disability, orientation or age; that infringes a patent, trademark, trade secret or copyright; that discloses others' confidential or private information; that encourages unlawful activity; or that contains malicious code
the requirement to warn meeting participants that they are being recorded is stated outright — for transcribing someone else's speech this is the key point
source, checked 2026-08-07
Legal
rights in what is uploaded stay with the user. Verbatim: «Customer retains all ownership rights in Customer Data», and the service is granted a licence «to access, process, copy, export, and display Customer Data, only as reasonably necessary (a) to provide, maintain and update the Services» and so on down a closed list. There is no separate clause in the terms about rights specifically in the transcript and translation, nor about commercial use of the result
the licence to the service is narrow and carries no right to train; but the word «transcript» is not used in the clause about rights — the conclusion is drawn from the definition of Customer Data
source, checked 2026-08-07
Catalogue section: all similar See also: catalogue index · scheduler comparison · find by situation · platform restrictions