Transcripts & summaries
How Eavesy transcribes sessions and structures the AI summary, and how to edit both.
Every recording produces two linked artifacts: the transcript (the complete record of what was said) and the summary (a structured distillation on top of it).
The transcript
- Built live with sub-second latency while you talk.
- Supports dozens of languages automatically.
- Fully editable — fix a name or tidy a line — and exportable as Markdown.
The transcript is always included on every plan. It’s never paywalled behind a summary-only tier.
Live transcription powers the Mac app and the in-app recorder. Uploaded audio files are transcribed with a high-accuracy batch model, so importing an existing recording produces the same transcript and brief.
The summary
When you stop recording, Eavesy generates:
- a title and TL;DR,
- decisions made in the session,
- action items with owners and due dates where mentioned, and
- open questions to follow up on.
It’s structured as clean Markdown, so you can copy the whole brief and paste it straight into Claude, Cursor, or ChatGPT — your agent gets the full context in one shot. More on feeding agents →
Editing and re-running
You can edit the summary directly, or re-generate it through a different template — for example, switching a session from a generic summary to a sales-call format. The underlying transcript never changes, so you never lose the source of truth.
Looking for an old decision across every session? Use semantic search to jump to the exact moment it was said.