How it works
Scout’s dictionary uses two layers working together: Layer 1: Speech engine hints — Your top dictionary entries are sent as hints to the speech engine before you start dictating. This helps the engine recognize your custom words during transcription — it’s more likely to hear “Kubernetes” if it knows you use that word. Layer 2: Fuzzy matching — After transcription, Scout runs a second pass on your machine. It compares words in the transcript against your entire dictionary. If a word is close enough to a dictionary entry, it’s automatically replaced. The matching is lenient enough to catch common transcription errors (like “post gress” → “PostgreSQL”) but strict enough to avoid false corrections on unrelated words.Adding words
- Open Settings → Dictionary
- Click Add word
- Type the correct spelling of the word (this is what Scout will output)
- Save
Examples
How matching works in detail
- Longest match first — Scout tries multi-word matches before single-word matches, so “Scout Voice” matches before just “Scout.”
- Multi-word entries — Dictionary entries with spaces can match across word boundaries in the transcript.
- Trailing punctuation preserved — If the original word had a trailing period or comma, it’s kept after replacement.
- Unlimited entries — Add as many words as you need. There’s no limit on dictionary size.