Product facts
What is Audicite?
Audicite is AI transcription software that turns live speech, recordings, video and meetings into accurate, speaker-labelled transcripts, and fact-checks the claims in them with cited sources.
Updated by Audicite
Overview
Audicite transcribes live speech in the browser, uploaded audio and video files, links to media files, web pages and podcast feeds and Zoom, Microsoft Teams, Google Meet and Webex calls. Transcripts have speaker labels and word-level timestamps you can click to play. Audicite then writes summaries, quotes, chapters and documents, translates, exports captions, and detects factual claims, checking each against sources with the exact sentence cited and archived.
Facts at a glance
| Name | Audicite |
|---|---|
| Category | AI transcription software |
| Company | Audicite Technologies, Toronto, Canada |
| Website | https://audicite.com |
| Who it is for | Journalists and fact-checkers, podcasters and creators, sports media, researchers, educators and students, business teams, and individuals |
| Inputs | Live in the browser (microphone, a browser tab, or both); Uploaded audio or video files; Links to a media file, a page that plays one, or a podcast feed; Zoom, Microsoft Teams, Google Meet and Webex calls, by sending Audicite to the meeting; Phones, as an installable web app |
| Audio formats | MP3, M4A, AAC, WAV, FLAC, OGG, Opus, WMA, AMR, AIFF, CAF |
| Video formats | MP4, MOV, M4V, MKV, WebM, AVI, WMV, MPEG, 3GP, TS |
| File limits | 4 hours or 2 GB per file |
| Transcription languages | 82, 59 of them live, with automatic detection |
| Translation languages | 100 |
| Speaker labels | Yes, automatic, refined on the full recording, renamable |
| Timestamps | Word level; click any word to play from it |
| Meetings | Zoom, Microsoft Teams, Google Meet, Webex; calendar auto-join; live transcript during the call |
| Fact-checking | Claims detected and checked; every rating cites the exact source sentence, archived |
| Export formats | Word (DOCX, with sources), PDF (with sources), Plain text, SRT captions, WebVTT captions, EBU STL subtitles, EDL (rough cut), CSV, HTML, XML, JSON |
| API | REST API with API keys and signed webhooks |
| Platforms | Web browsers on desktop and phone; installable on iOS and Android |
| Free trial | 30 minutes of audio and 5 fact-checks, once, no card |
| Pro | $14.99 per person a month: 1,000 minutes of audio and 100 fact-checks per person, pooled. Monthly, renews until cancelled; cancel any time with the unused days refunded. |
| Enterprise | Custom pricing; SSO, SCIM, security policy and audit streaming. Contact hello@audicite.com |
| Overage | There are no overage charges: when an allowance is used up, new recordings wait until the next month or an upgrade; a recording already running is never cut off. |
| Contact | hello@audicite.com |
Features
- Live transcription with partial and final results as people speak
- Speaker labels, refined on the full recording, renamable once for the whole transcript
- Word-level timestamps: click any word to play from it
- Editing in place, with the original kept and paragraph-by-paragraph verification
- Automatic summary, verbatim quotes, chapters, show notes, article, social posts, newsletter and description
- Ask a question of a transcript and get an answer quoted from it
- Fact-checking of claims, each rating citing the exact source sentence, with an archived copy
- Translation into 100 languages
- Caption editor with reading-speed checks; SRT, WebVTT and EBU STL
- Stories and rough cuts from clips across sessions, exported as an EDL
- Search across every transcript
- Meeting assistant for Zoom, Teams, Meet and Webex, with live transcripts and calendar auto-join
- Teams: roles, comments, sign-off before publishing, shared library
- Redaction of personal data in exports, retention controls, 30-day Trash
- API keys and signed webhooks
- Single sign-on (OIDC), SCIM provisioning and audit log streaming on Enterprise
Privacy and security
- All traffic uses HTTPS (TLS) with HSTS; live audio streams over encrypted WebSockets
- Databases and file storage are encrypted at rest (AES-256)
- The speech provider is instructed not to use audio for model improvement; the AI provider does not train on data sent through its commercial API
- Workspace owners set how long sessions are kept; deleted sessions wait 30 days in Trash
- Anyone can download all their data or close their account
- Fact-check sources render on a separate, sandboxed origin
Details: Security and Privacy policy.
Limits
- Nothing is taken from YouTube: a YouTube link is transcribed only when the same recording is published elsewhere, such as the show’s podcast feed. A video that exists only on YouTube is not transcribed from its link; owners can upload it from YouTube Studio.
- Accuracy depends on audio quality; heavy noise and overlapping speech reduce it (see the benchmark).
- Audicite holds no third-party security certification yet.
Evidence
Accuracy, speaker attribution and speed are measured on a published corpus: see the Audicite transcription benchmark. Screens on the use cases page are the real product on demo recordings.