Audio Briefings

Cortexa generates podcast-style audio briefings from your recent captures — a spoken synthesis of your latest thoughts, decisions, and TODOs.

Setup

1. Settings → Briefings → TTS settings → enable TTS
2. Choose a TTS provider:
   • OpenAI (cloud) — paste your OpenAI API key. Pick a voice: alloy, echo, fable, onyx, nova, or shimmer
   • Piper (local, bundled) — no API key needed. Runs entirely on-device for offline/private use.
3. Save

Generating a Briefing

1. Set the "Recent items" count (e.g., 7 items)
2. Click Generate
3. The job goes through: scripting → voicing → done

The pipeline:

1. Script — LLM reads your N most recent captures and writes a 2-4 minute spoken briefing script in second person ("You captured a decision about...")
2. Voice — OpenAI TTS or Piper (local) synthesizes the script into audio (mp3)
3. Store — audio is encrypted and stored locally

Playback

Once done, click Load audio on the briefing card. A waveform audio player appears with:

• Play/pause controls
• Frequency visualizer (real-time FFT)
• Seek bar
• Duration display

Script Without Audio

If TTS is disabled, you still get the script — click the Script disclosure to read it. Useful if you don't need audio but want a synthesis of recent activity.

Scheduled Briefings

In Settings → Schedules, enable the Daily briefing to auto-generate at a specific time (e.g., 8:00 AM). The briefing is queued automatically by the background ticker.

Managing Briefings

• Delete — click the trash icon on any briefing card to remove it (deletes the encrypted audio file from disk)
• Briefings accumulate over time; delete old ones to reclaim disk space

Scopes

Scope	Description
Recent	Last N items by creation date (default)
Item	Specific item IDs (backend supported, UI coming)