talat

The application captures audio from any conferencing software, records both participants, and performs real‑time transcription using an on‑device AI model. Transcribed segments are attributed to speakers and can be edited, split, or reassigned during the meeting without interrupting the session. After the call ends, a local language model produces a concise summary that highlights key points, decisions, and action items. All data—including recordings, transcripts, and notes—is stored locally on the user’s computer and never uploaded to the cloud.

Designed for users who need private, searchable meeting records, the tool works with common platforms such as Zoom, Teams, and Google Meet. It runs in the background, detects microphone usage, and provides brief notifications while keeping focus on the primary meeting application. Users can export notes to folders, trigger webhooks, or integrate their own language models via customizable system prompts.

The software is offered as a one‑time purchase with a limited free tier and requires no account or subscription. It emphasizes local processing, data ownership, and compatibility with existing meeting workflows, allowing it to be used alongside other note‑taking solutions without replacing them.