> ## Documentation Index
> Fetch the complete documentation index at: https://docs.openwhispr.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Custom dictionary

> Improve transcription accuracy for names, jargon, and technical terms.

Add words to the custom dictionary so the speech recognition model recognizes them correctly.

## How it works

Words you add are provided as context hints to the transcription model. This makes them more likely to be recognized accurately. Works with both local Whisper and cloud transcription.

## Adding words

1. Open **Dictionary** from the sidebar
2. Type a word or phrase and press Enter
3. Words are saved immediately and used on your next transcription

## What to add

* Uncommon names — "Sergey", "Xanthe", "Priya"
* Technical jargon — "Kubernetes", "OAuth", "PostgreSQL"
* Brand names — "OpenWhispr", "whisper.cpp", "sherpa-onnx"
* Domain-specific terms — "amortization", "polymerase", "arbitrage"
* Internal terms — project names, product codenames, team names

## Auto-learn

OpenWhispr can detect when you correct a transcription and automatically suggest words for your dictionary. This learns from your editing patterns over time.