Transcripto

Live bilingual transcript

Listen in one language, read in another.

Transcripto captures live speech from your microphone, transcribes it in the source language, translates it into your target language, and keeps a readable scrolling transcript you can reopen later.

OpenAI transcription Faithful translation Installable PWA Local session history

OpenAI API key

Stored locally in this browser only. Use a personal key you control.

Source language Target language

Live transcription model Use the new realtime Whisper option for your next test run, or keep the current balanced default. Glossary / context (optional) Used as context to improve transcription and translation quality. Speaker names (optional) Used for manual speaker timing and to name speakers after the final saved-audio analysis pass.

Preferred microphone Uses the default microphone when none is selected. Device names usually appear after microphone permission is granted.

Echo cancellation Noise suppression Auto gain control

Use headphones if the room audio is loud or echoey.
Keep the phone close to the speaker or PA source.
Add names, brands, acronyms, and product terms to the glossary.
Add speaker names up front if you already know who will talk.
Stop whenever you do not need capture, then Resume the same session when you are ready.
The last active session can be resumed after refresh or reopening.

Architecture: this build is static GitHub Pages, so requests go directly from the browser to OpenAI using your own locally stored API key.

Current session

No active session

Start a session to begin.

Duration 00:00

Speech-only 00:00

Rows shown 0 0 saved segments

Status: idle. Waiting to start.

🎙️ Manual speaker timing

00:00.00

Current speaker

Insights

Speaker analysis & session audio

No speaker analysis data yet.

Whole session audio Local session recording will appear here while capture runs.

Session behavior

Stop closes the live connection and stops microphone sending for the current session.
Final transcript segments are saved locally after each completed turn.
After long listening sessions, the connection rolls over automatically.
Resume continues the same logical session and transcript history.
Final speaker analysis runs after Stop using the saved session audio, not during live capture.

OpenAI API key

Default source language Default target language

Live transcription model This changes the live speech-to-text model used the next time capture starts. Text size Theme Auto-follow live reader Timestamp style

Preferred microphone Uses the default microphone when none is selected. Stop and resume capture to apply changes during an active session.

Echo cancellation Noise suppression Auto gain control

Glossary / context Default speaker names Used as the default speaker-name order for new sessions and can also refresh labels in the current session.

Deleting local session data removes transcript history from this browser only. It does not affect GitHub Pages deployment.

Transcripto

Listen in one language, read in another.

Start a session

Mic settings

Before a keynote or meeting

No active session

Capture controls

Live bilingual transcript

Speaker analysis & session audio

Previous sessions

Settings

Mic settings

Manage browser storage