Custom Vocabulary

Add domain-specific terms, names, and technical jargon to improve transcription accuracy.

Why Use Custom Vocabulary?

Wispli's speech recognition is trained on general language. It might not recognize:

• Technical terms: Kubernetes, GraphQL, PostgreSQL
• Product names: Your company's product names
• Acronyms: Internal acronyms and abbreviations
• Names: Uncommon names of people or places

Adding Terms

Via Settings

1. Open Settings (gear icon)
2. Go to Vocabulary tab
3. Click Add Term
4. Enter the term exactly as you want it spelled
5. Click Save

Bulk Import

Import multiple terms from a text file:

1. Create a .txt file with one term per line
2. In Settings → Vocabulary, click Import
3. Select your file

Example file:

Kubernetes
GraphQL
PostgreSQL
TailwindCSS
NextJS

How It Works

When you speak, Wispli:

1. Transcribes your speech
2. Checks for phonetic matches to your vocabulary
3. Replaces recognized words with correct spelling

Example:

Without vocabulary:
  "We deployed to Cooper Netties" ❌

With "Kubernetes" in vocabulary:
  "We deployed to Kubernetes" ✅

Best Practices

Do Add

• ✅ Technical terms specific to your field
• ✅ Product and brand names
• ✅ Uncommon proper nouns
• ✅ Industry acronyms

Don't Add

• ❌ Common English words
• ❌ Words Wispli already recognizes correctly
• ❌ Misspellings

Phonetic Matching

Wispli uses intelligent phonetic matching:

You Say	Without Vocab	With Vocab
"supa base"	"supa base"	"Supabase"
"next jay ess"	"next JS"	"NextJS"
"post gres"	"post gres"	"PostgreSQL"

Managing Vocabulary

Edit Terms

1. Go to Settings → Vocabulary
2. Click the edit icon next to a term
3. Make changes and save

Delete Terms

1. Go to Settings → Vocabulary
2. Click the delete icon next to a term

Export

Export your vocabulary for backup or sharing:

1. Settings → Vocabulary → Export
2. Save the .txt file

Team Vocabularies (Coming Soon)

Share vocabulary lists across your team for consistent recognition of company-specific terms.

Next Steps

• Shortcuts - Quick text replacements
• Voice-to-Text - Transcription basics