Podcast & interview transcripts
Transcribe hour-long episodes, get speaker-split segments, and paste the text into your show notes — ready for SEO and accessibility in one click.
AI Speech to Text
Transcribe audio to text with 99% accuracy
Upload or record audio and get a clean, accurate transcript in seconds. Powered by OpenAI Whisper — supports 90+ languages and speaker diarization.
Record
00:00
Tap to start recording
How it works
Transcribe audio in three simple steps
No setup, no account needed to try. Upload your audio and get a clean transcript in seconds.
01
Upload or record audio
Drag and drop an audio file — MP3, WAV, M4A, WEBM, or MP4 — or tap the mic to record directly in your browser. Up to 25 MB supported.
02
Hit Transcribe
Lyssna sends your audio to OpenAI Whisper. It auto-detects the language, handles accents and code-mixed speech, and returns a clean, punctuated transcript in seconds.
03
Copy, download, or re-voice
Copy the text, download it as a .txt file, or paste it straight into the TTS studio to re-voice it in any language — without leaving the app.
Every kind of audio
One transcriber. Six creator workflows.
Reels, TikTok & YouTube captions
Drop in your vertical video, pull the transcript, and generate burn-in captions. Boost watch time without re-typing a word.
Voice notes to clean text
Speak your thoughts. Lyssna punctuates, paragraphs, and hands you a tidy note ready for Notion, Docs, or your Substack draft.
Meeting & call transcription
Upload Zoom or Google Meet exports, get diarized transcripts, and share decisions without paying an extra meeting-bot subscription.
Lectures & study audio
Turn class recordings into searchable, summarizable text. Review at 2× the speed, highlight key moments, and export as notes.
Accessibility & closed captions
Make audio and video content work for deaf and hard-of-hearing audiences. Accurate Whisper output, multi-language support, compliant with WCAG captioning goals.
Questions, answered
Everything about our transcriber
What is speech-to-text and how accurate is Lyssna?
Speech-to-text (STT) converts recorded or live audio into written text using AI. Lyssna runs OpenAI’s Whisper model, which routinely hits 95%+ accuracy on clear English and strong multilingual performance. Noisy calls, heavy accents, and cross-talk all reduce accuracy — the playground on this page lets you test a real clip before committing.
Is your speech-to-text tool free to use?
Yes. The playground on this page transcribes up to a 25 MB clip without signup. Signed-in accounts receive starter credits, unlock history, longer files, and the full re-voice workflow.
Which languages do you support?
Whisper covers 90+ languages with automatic detection — including English, Hindi, Tamil, Telugu, Bengali, Marathi, Spanish, French, German, Portuguese, Japanese, Korean, and Chinese. We also handle code-mixed speech (for example, Hinglish) without requiring you to pick a language manually.
Can I transcribe video files, not just audio?
Yes. Drop in MP4, MOV, WebM, and most common video containers. We extract the audio track and transcribe it — no need to convert first.
Does it identify different speakers?
Yes. Our backend returns diarization segments whenever they’re available — speaker labels, start/end timestamps, and per-sentence chunks. Perfect for interviews, panels, and multi-person meetings.
How long can my audio file be?
The free playground accepts files up to 25 MB (roughly 25 minutes at standard podcast quality). Paid plans raise this limit and support background batch uploads for longer episodes.
Is my audio private?
Playground uploads are processed in-memory and deleted once the transcript is returned. Signed-in accounts get encrypted storage with a clear-at-any-time button — you own every file.
Can I convert the transcript back into a different voice?
That’s the Lyssna difference. A single click pushes your transcript into our TTS tool, where you can re-voice it in ElevenLabs, Inworld, or MiniMax — ideal for localization, voice swaps, and creator remixing.