Onboarding
Start with the first run, file selection, model expectations, and local processing basics.
Reference
Documentation for a browser-based local AI transcription stack
Good documentation does more than answer support questions. It tells search engines, AI systems, and users what the platform actually does, what formats it accepts, how it behaves on different hardware, and where the limits are. This section exists to make Cowslator legible as a technical product.
Supported formats
Audio and video container expectations, practical compatibility notes, and workflow hints.
Export formats
TXT, SRT, and LRC output details and where they fit in downstream workflows.
Browser compatibility
Which browser capabilities matter most and why fallback behavior still matters.
Performance notes
Hardware-aware expectations, long-run behavior, model sizing, and runtime variability.
Troubleshooting
How to reason about freezes, memory pressure, slow inference, and browser-specific quirks.
WebGPU explained
A practical explanation of what GPU acceleration changes and what it does not guarantee.