Free Voice Note to Text — AI Transcription & Notes
Upload voice memos, voice notes, and audio recordings. Get accurate transcripts with chapters, topics, AI-generated summaries, and searchable playback. Copy key takeaways into Notion, Apple Notes, Obsidian, or any note-taking app.
AI Q&A
Chapters & Topics
Speakers
50+ Languages
Private & Encrypted
| Model | Speakers | Timestamps | Price | WER |
|---|---|---|---|---|
Qwen FlashDefault | Word (10 langs) | $1.05/hr ($0.0176/min) | – | |
Voxtral Mini | Word (13 langs) | $1.12/hr ($0.0187/min) | – | |
GPT-4o | Sentence | $3.23/hr ($0.0538/min) | – |
Qwen FlashDefault
$1.05/hr No speakers|Word (10 langs)|WER –
Voxtral Mini
$1.12/hr Speakers|Word (13 langs)|WER –
GPT-4o
$3.23/hr Speakers|Sentence|WER –
View all supported languages
Your transcriptions
No transcriptions yet. Start by pasting a link or uploading a file above.
Try your first transcription free
Free credit included — no credit card required. Transcribe video or audio in 50+ languages. Pay only for what you use after that.
FAQ
Frequently asked questions
Upload your voice memo or audio recording to Transcribe.so. Free credits are included on sign-up — no credit card required. You get an accurate transcript with timestamps, chapters, topics, and AI Q&A.
Transcribe.so supports MP3, WAV, M4A, AAC, FLAC, OGG, MP4, MOV, and WebM. Voice memos from iPhone, Android, and any recording app work out of the box.
Yes. Every transcription includes auto-generated chapters, topic detection, and AI Q&A. Ask a question about any recording and get a cited answer linked to the exact second.
Yes. The full transcript with timestamps and chapters is available as plain text. Copy it directly into Notion, Apple Notes, Obsidian, Google Keep, OneNote, or any note-taking app.
Yes. All uploads are encrypted, stored privately, and can be deleted at any time — instant removal. Transcribe.so uses Cloudflare and OpenAI infrastructure with no third-party data sharing.
Accuracy depends on the speech-to-text model and language. Transcribe.so lets you pick the best model for your language — GPT-4o Transcribe for speaker-labeled interviews, Qwen3-ASR-Flash for word-level precision, or Voxtral for cost-sensitive long recordings.
Try your first transcription free
Free credit included. No credit card required. Transcribe video or audio in seconds.