Voice to Text Online (V2T) is a browser dictation and transcription service. Browser dictation works with no account and the text stays in local storage — 'The text is not sent to V2T' (account-free dictation uses the browser's own speech recognition, so audio may go to the browser vendor); signed-in live transcription streams mic audio to OpenAI. Uploading audio or video instead sends files for cloud transcription with speaker labels, an AI Enhance pass that cleans punctuation, grammar, casing and paragraphs, and generated summaries, email drafts, action items and translations.
Around the core there is a Chrome extension for dictating into Gmail and other sites, YouTube URL to text conversion, text-to-speech with MP3 download, Google Drive and Dropbox imports, a developer API and an MCP endpoint (currently YouTube transcripts and TTS only; Zapier, OneDrive, Notion, Slack and LinkedIn integrations are announced but not live).
Pricing is the most granular of this batch, verified on the page: browser dictation is free; a free account includes 5 uploaded-file minutes and 20 live minutes once per account; Pay as you go is $5 for 100 file minutes that never expire; Starter is $7/month (200 file + 60 live minutes, $59/year); Pro is $15/month (600 file + 200 live minutes, $149/year); and Live PAYG sells 20 live minutes for $1 with a $5 minimum on a separate non-expiring balance.
Limits and privacy: the free account allowance is one-time only; MCP does not take file transcription yet. The privacy policy (November 12, 2026) discloses that live audio streams encrypted to OpenAI (recordings not kept), uploads go to AssemblyAI on its free tier with model improvement enabled — so ordinary uploads may help train the upstream model (new Drive imports use an EU route excluded from training) — plus Anthropic for enhancement and Google Cloud TTS; no self-service deletion (aim: 7 business days).