Free
Every feature on the free plan. No credit card to start.
The fastest, most accurate translation engine
BufferMind runs on Soniox, the speech engine that tops independent accuracy benchmarks. Every word hits the screen as it's spoken, and the translation lands about a second behind — through accents, fast talkers, and mixed languages.
Powered by Claude
Andy, your AI tutor, runs on Claude Sonnet 5 and Opus 5 — the most capable models Anthropic has ever shipped. Summaries, answers, and study notes with real understanding behind them.
Simultaneous interpretation
Put on your headphones and the translation is spoken to you, live — only you hear it, never the room. Like carrying a personal interpreter.
Synced to the cloud, on every device
Record on your phone, pick it up on your Mac or the web — every session syncs automatically. Audio uploads while you're still recording, so a dead battery or a crash can't take your recording away.
Hold to talk, translated both ways
For face-to-face moments: each side holds a button and speaks their own language — the translation appears and speaks out loud instantly. Travel, small talk, real conversations.
Better with classmates
Chat, call, and share recordings with friends — send a whole lecture to a classmate, swap notes, review together.
Free. Private.
On every screen.
One app, three platforms. Every feature unlocked. Your data never used to train models.
Your data stays yours
Mic activates only when you tap Record. Delete anytime. Never used for AI training.
iPhone · iPad · Mac · Web
Record on one device, review on another. Everything syncs automatically.
FAQ
Is it really free? What does it cost?
Yes. The free plan includes every feature with a monthly allowance of about 3 hours of transcription (shared with AI questions and summaries), and no credit card to start. Pro raises it to about 25 hours a month for $9.99/month or $19.99 per 3-month term; Max offers about 80 hours for $19.99/month or $44.99 per term. Cancel anytime from inside the app. Connect with your first friend in BufferMind and you both get a free month of Pro.
What languages are supported?
60+ languages are auto-detected for transcription and translation — including English, Chinese (Mandarin & Cantonese), Spanish, Hindi, Korean, Japanese, French, German, Arabic, Russian, and more — powered by Soniox's unified model. The BufferMind app interface is currently localized in 6 languages.
Which devices can I use it on? Is there a Windows or Android app?
Mac, iPhone, iPad, and any modern browser. The Mac app needs macOS 14 or later on Apple Silicon (M1 and newer); the iPhone and iPad app needs iOS 16.2 or later and an App Store account outside mainland China. There is no native Windows or Android app — open BufferMind in your browser there instead. Sign in with the same account and your sessions sync automatically across every device.
Can it hear online meetings and videos, not just my own voice?
Yes, on Mac and in the browser. The Mac app captures the audio playing on your computer directly, so a Zoom or Google Meet call, a lecture recording or any video is transcribed without going through the microphone. In a Chromium-based browser (Chrome, Edge) you pick the meeting tab and turn on “Also share tab audio”. On iPhone and iPad, iOS does not let apps capture other apps' audio, so BufferMind listens through the microphone — put the call on speaker and it works.
Is my audio stored? Is it used to train AI?
No. We never use your recordings to train AI models. The microphone turns on only when you tap Record, never in the background. Audio is streamed to our speech-to-text partner (Soniox) only during transcription and is not retained afterward. Cloud copies of your sessions are tied to your account, encrypted, and can be deleted anytime from Session History.
Do I need an internet connection?
Yes. Real-time transcription and translation stream audio to the cloud, so a stable internet connection is required. We're exploring on-device modes as lightweight models mature.
What happens when my monthly allowance runs out?
Recording, AI questions and summaries all draw on one allowance that resets on the 1st of each month — the sidebar shows how much is left. When it runs out, everything you have already recorded stays exactly where it is; you just can't start new recordings or ask Andy more until the reset, or until you move up to Pro or Max.
How well does it handle accents and technical terms?
Transcription runs on Soniox's real-time model, built for accented speech and for speakers who switch between languages mid-sentence. For course-specific vocabulary — names, acronyms, formulas — paste your syllabus, notes or glossary into that course's Knowledge Base, and BufferMind uses it as context for its summaries and for Andy's answers.
What can I do with a recording after class?
Every session is saved with the audio, transcript and translation kept in sync — tap any line to jump back to that moment. BufferMind writes a title and a summary for each session, groups them by course, and Andy can answer questions about what was said. Transcripts can be copied or exported as text to share with classmates.
Can BufferMind tell different speakers apart?
Yes. Enable speaker identification and BufferMind labels “Speaker 1,” “Speaker 2,” and so on within a session. Voiceprints are processed only during that session by Soniox and are never stored by us.
