About

pronounce.how is an open-source English pronunciation guide that provides IPA transcriptions, audio, and syllable breakdowns for 14,231 words across four major English dialects: American, British, Canadian, and Australian.

Why This Exists

Most pronunciation resources focus on a single dialect or lack the transparency to show where their data comes from. pronounce.how is built on the belief that pronunciation data should be open, verifiable, and comprehensive across regions. Every entry shows its source, confidence level, and whether it's been verified by native speakers.

Data Sources

📘
CMU Pronouncing Dictionary

~134k American English pronunciations in ARPAbet notation. The gold standard for US pronunciation data.

📗
Britfone

~16k British English pronunciations in IPA. Curated phonetic dictionary for Received Pronunciation.

🔊
eSpeak NG

Open-source speech synthesizer used for Australian English pronunciation and gap-filling.

📖
Wiktionary (via Wiktextract)

Cross-validation source for IPA transcriptions and part-of-speech data.

How Confidence Works

Every pronunciation variant has a confidence score from 0 to 1. This reflects how trustworthy we believe the transcription to be:

  • 90%+Primary dictionary source (CMU Dict, Britfone), confirmed by cross-validation
  • 60-89%Secondary source, derived variant, or partial cross-validation
  • <60%eSpeak gap-fill or unverified derivation. Help us verify these!

Open Source

All pronunciation data is open source and available on GitHub under CC BY-SA 4.0. The website code is also open source. We welcome contributions — especially native speaker verifications for Canadian and Australian English.

License

Pronunciation data: CC BY-SA 4.0. Website code: MIT. Please see individual source attributions in the data repository.