SME Trans

Transcription Engine

Model Size

Language

Runs entirely in your browser — no audio is uploaded anywhere, and no API key is needed. The first time you use it, the model is downloaded and cached in your browser (faster on future uses).

Drag your MP4/audio here, or click to browse

Any size — handled automatically

0:00

Include meeting/tab audio (capture other speakers too)

Record straight from your microphone — handy if you're sitting in on a live Teams call and want to capture the conversation.

Want each speaker labeled automatically? Switch to the Cloud engine in Settings, connect Deepgram or AssemblyAI, and keep "Identify speakers" turned on.

Audio/Video URL or YouTube Link

YouTube links are supported. By default, captions are fetched automatically (no download needed). If AssemblyAI is your active Cloud provider, the audio is pulled instead and run through real transcription (so speaker labels work too). For other links, the server must allow cross-origin access (CORS). Private Teams/SharePoint links won't work here.

Knowledge Document

Speech language

Answer automatically after each sentence Also transcribe my own mic (shown as "You", never answered)

Audio from the shared tab is streamed to Deepgram for live transcription, and each question plus the matching parts of the document are sent to your Groq or OpenAI key to get the answer. The document itself is stored only in this browser.

Ready to Transcribe

—

API Keys — Protected