Capabilities

Inline custom pronunciation

Control how a word sounds in one request by replacing it with inline IPA.

Inline custom pronunciation controls how a word sounds in one synthesis request by replacing the written word with its International Phonetic Alphabet (IPA) pronunciation. It is useful for generated text, one-off corrections, and applications that want to choose pronunciations per request.

Add inline IPA

Inworld TTS supports inline IPA phoneme notation for custom pronunciation. Use the International Phonetic Alphabet (IPA) format, wrapped in slashes (/ /).

Use standard English IPA phonemes only. Do not use CMU dictionary phonemes (ARPAbet) or phoneme sets from other languages. Inworld TTS is optimized for English IPA symbols.

For example:

Suppose you are building an AI travel agent, and it is recommending the destination Crete, which is pronounced /kriːt/ ("kreet") in English.
You can ensure the correct pronunciation by passing it inline: Your interests are a perfect match for a honeymoon in /kriːt/.

Replace one word at a time. Wrap only a single word in slashes for each substitution. Do not wrap multiple words or phrases within the same slash delimiters.

Correct: Welcome to /kriːt/, the beautiful island.
Incorrect: Welcome to /kriːt ðə bjuːtɪfəl aɪlənd/.

The model will substitute the IPA pronunciation wherever it appears inline in your text. If the text is generated by an LLM, you can simply replace the original spelling with the IPA transcription before passing it to the TTS model.

Find the right IPA phonemes

If you are unsure of the correct phonemes, there are several ways to find them:

Ask an LLM like ChatGPT: For example, you can ask:

"What are the IPA phonemes for the word Crete, pronounced like 'kreet'?"
Use reference websites:
- toPhonetics is a free tool that converts English text to IPA transcription.
- Vocabulary.com's IPA Pronunciation Guide provides tables of IPA symbols with example words.

Once you have the correct phonemes, you can embed them directly into your TTS request: Your adventure in /kriːt/ begins today.

When to use inline pronunciation

Use inline pronunciation when the correction should apply to one request, when an application already transforms generated text, or when the same spelling must sound differently in different contexts.