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

# Custom Pronunciation

> Control how Plivo Text-to-Speech says brand names, acronyms and jargon with inline pronunciations or saved lexicons.

> [!NOTE]
>
> Fix how a word is spoken. Pass pronunciations inline for one request, or save them in a lexicon and reuse it by ID.

## Two ways to set a pronunciation

| | Inline `pronunciations` | Saved lexicon |
| - | - | - |
| Where | In each batch request or realtime `text` message | Created once with the Lexicons API |
| Reuse | No | Pass `lexicon_id` on any request |
| Limit | 100 per request | 100 lexicons per org, 100 rules each |

## Pronunciation kinds

| `kind` | Write the value as | Example for "Plivo" |
| - | - | - |
| `respell` | Plain-English respelling | `plee-voh` |
| `ipa` | International Phonetic Alphabet | `ˈpliːvoʊ` |
| `arpa` | ARPAbet | TBD |

## Inline example

```json theme={null}
{
  "voice_id": "<voice_id>",
  "text": "Welcome to Plivo",
  "pronunciations": [{"word": "Plivo", "value": "plee-voh", "kind": "respell"}]
}
```

`kind` is required on each inline entry.

## Lexicon example

```bash theme={null}
curl -X POST "https://api.plivo.com/v1/Account/<auth_id>/TTS/Lexicons/" \
  -u '<auth_id>:<auth_token>' \
  -H "Content-Type: application/json" \
  -d '{"name": "brands", "kind": "respell", "rules": [{"word": "Plivo", "value": "plee-voh"}]}'
```

The reply has a `lexicon_id` (`lx_...`). Pass it as `lexicon_id` when you synthesize.

## Related

* [Lexicons API](https://plivo.com/docs/models/text-to-speech/api/lexicons.md)
* [Batch API](https://plivo.com/docs/models/text-to-speech/api/batch.md)
