Navigate to yapr.lol and hit 'Start Free'. Sign in with Google or GitHub for the fastest ride — or create an account with email and password if that's more your style. Got a passkey? Even better — one touch and you're in. The whole flow takes under 15 seconds.
Your account is personal and private. We keep only what's needed to sign you in — nothing extra, nothing creepy. No data is shared or sold. Account deletion wipes everything clean.
Hit the mic button or press Space. Recording starts immediately. Speak at any pace, in any language — pause, breathe, switch topics. The AI handles natural speech, not scripted input.
When you're done, press again to stop. Audio is instantly encoded as WebM/Opus in-browser, then sent over TLS 1.3 + AES-256 encryption to our servers. Nothing is stored locally at any point.
Your audio is staged briefly on Cloudflare's global edge network, then routed to dedicated GPU inference endpoints. We use state-of-the-art speech AI models trained on 680,000+ hours of real-world multilingual audio.
The models natively understand 100+ languages — detecting them automatically, no settings required. You can switch languages mid-sentence, mid-recording. The model keeps up. Median round-trip: 1.8 seconds.
Transcribed text appears on screen, formatted and ready. Punctuation is auto-added from natural speech rhythm. The output is clean, readable prose — not fragmented tokens or raw phonemes.
A metadata bar shows duration, word count, processing latency, and detected language. If history is enabled, the transcript is automatically saved and searchable indefinitely. History is off by default — enable it in Settings to save transcripts.
One tap copies the full transcript to clipboard — paste it anywhere: Slack, a doc, a CRM, a code editor. No file to download, no app to open.
Your History page saves every transcript: searchable, full-text, with date, duration, and language. Record, copy, record, copy — Yapr is built for rapid iteration at whatever pace your workflow demands.
The moment transcription completes, the staged audio file is deleted from our servers automatically and irrevocably. There is no way to retrieve it — not even by us.
A TTL (Time-To-Live) policy on the staging layer guarantees deletion within 60 seconds even if transcription fails. This is not policy — it's architecture. Without a long-term storage layer, retention is structurally impossible.
Everything your workflow needs. Tap any option to learn more.
We didn't write a privacy policy and call it done. We built a system that structurally cannot retain your audio — because without a long-term storage layer, retention is impossible by design.
< 60s
Maximum audio retention time
TLS 1.3
Encryption standard in transit
AES-256-GCM
Encryption at rest (transcripts)
OAuth 2.0
Auth method — no passwords stored
HSTS + CSP
Security headers enforced on all routes
GDPR + CCPA
Full regulatory compliance
0 bytes
Audio retained after transcription
No credit card. No trial period. No 'freemium' tricks. You get full access from the moment you sign in — completely, genuinely free.
No credit card, ever
No download or install
No phone number required
No ads, no tracking
No hidden limits
No trial that expires
Six steps. Under two seconds. TLS 1.3 + AES-256. Your audio deleted permanently. Start free.
