Spoken terms lets users create lists of words (e.g., names, acronyms, company-specific terms) and apply them during a meeting to improve transcription accuracy. Terms are checked during the speech-to-text step before translation happens, correcting recognition of customer-specific language.
Spoken terms can be used alongside glossaries to give you full control over the accuracy of your multilingual content. See the table below to understand the purpose of each feature and at which stage they are applied.
| Spoken terms | Glossaries | |
| When | Transcription | Translation |
| Purpose | Improve how words are recognized | Control how words are translated |
Availability
Spoken terms is a customization feature available for online (see DeepL Voice for Meetings) and in person meetings (see DeepL Voice for Conversations). This feature is not available with the DeepL Voice for Meetings Core plan or for group conversations.
You can create Spoken terms lists for all languages transcribed by DeepL. It is not available for languages transcribed by a third-party provider. Learn more: DeepL Voice languages.
Character limits
- 300 characters per language, per list
- Languages at the character limit are greyed out
- Terms applied to All languages count toward each individual language's limit separately