Fireflies Talk Adds AI Dictation Across Mac and Windows Apps
Fireflies Talk brings systemwide AI dictation to Mac and Windows, with local transcripts, no saved recordings and access through existing subscriptions.
Summary
On September 29, 2026, Fireflies.ai expanded beyond meeting notes with Fireflies Talk, a dictation feature in its Mac and Windows apps. Pressing Fn on Mac or Ctrl + Win on Windows starts recording from any open app, then inserts a cleaned, neatly formatted transcription into the active text field. CEO Krish Ramineni said people speak roughly 3x faster than they type and positioned Talk as another step toward Fred, an AI teammate spanning email, Slack, documents, prompts and other work.
Talk is bundled into existing Fireflies subscriptions rather than sold separately. Fireflies says it retains no voice recordings and stores dictation transcripts locally. The feature enters a field occupied by Wispr Flow, Willow and SuperWhisper. Speechify has also added cross-app dictation, while Wispr took the reverse bundling approach by giving its meeting notetaker free to existing dictation users. Fireflies plans a context-aware assistant that uses meeting transcriptions to automate workflows. It introduced meeting insight mini-apps in 2025 and added inbox management and email drafting in August 2026, but details on the planned assistant, dictation accuracy and language support remain limited.
Positives
- Fn on Mac and Ctrl + Win on Windows activate Fireflies Talk from any open application.
- Existing Fireflies subscriptions include Talk without a separate charge.
- Voice recordings are not retained, while dictation transcripts are stored locally.
- People speak roughly 3x faster than they type, supporting Fireflies’ push toward voice-based work.
- Fireflies added inbox management and email drafting in August 2026 after launching meeting insight mini-apps in 2025.
Risks & concerns
- Wispr Flow, Willow, SuperWhisper and Speechify already compete in cross-app AI dictation.
- Fireflies’ context-aware workflow assistant remains planned rather than available.
- Dictation accuracy, supported languages and detailed rollout terms remain unspecified.