Agent skill · marketing growth · jeremylongshore
granola-performance-tuning
Optimize Granola transcription accuracy, note quality, and processing
Why this skill is useful
Provides specific audio setup recommendations and meeting practices that enhance Granola's transcription accuracy and processing speed.
What it needs
About 4k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
Granola Performance Tuning Overview Optimize Granola output quality across three dimensions: audio/transcription accuracy, AI enhancement quality, and integration speed. Granola's AI (GPT-4o/Claude) produces better output when it has clean audio, well-typed notes, and structured templates. Prerequisites Working Granola installation with meetings captured Willingness to improve audio setup and meeting practices At least 3-5 meetings captured to establish baseline quality Instructions Step 1 — Optimize Audio for Transcription Granola captures system audio from your device. Transcription accuracy depends entirely on audio quality: Hardware recommendations (by priority): Setup Accuracy Impact Recommendation ------- ---------------- ---------------- Wired headset with mic Highest Best for solo/remote meetings USB condenser mic High Best for in-office, multiple speakers Laptop built-in mic Medium Acceptable for quiet environments Bluetooth headset Variable May cause dropouts — test first Speakerphone in room Low Echo and distance degrade accuracy Audio configuration checklist: [ ] Correct input device selected in System Settings > Sound > Input [ ] Input volume at 75-100% (not too low, not clipping) [ ] Audio enhancements disabled (Windows: right-click device > Properties > disable enhancements) [ ] No conflicting virtual audio software (Loopback, BlackHole, etc.) [ ] Bluetooth device stable (or switch to wired if experiencing drops) Room setup: [ ] Minimal background noise (close doors, turn off fans) [ ] Soft surfaces to reduce echo (avoid glass-walled conference rooms) [ ] Mic within 12 inches of speaker(s) [ ] Meeting participants using headsets (reduces echo and crosstalk) Step 2 — Improve Meeting Practices These behaviors directly improve Granola's output: Practice Impact Why It Helps ---------- -------- ------------- State names when assigning work High "Sarah, can you handle the API spec?" enables correct attribution Use explicit action language High "Action item: …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/granola-performance-tuning--0b1427