Rosette's Language Identification API supports detection across 364 language-encoding-script combinations, covering over 66 distinct languages. The agent sends the customer's text to the endpoint and receives the detected language with a confidence score, enabling accurate routing and response language selection.

The agent sends two name objects (Person, Location, or Organization type) to Rosette's Name Similarity API. Rosette analyzes the names across linguistic variations, transliterations, and cultural naming conventions, then returns a similarity score between 0 and 1. You define the threshold for a match based on your verification requirements.

Yes. Rosette's Address Similarity endpoint compares addresses provided as single strings or structured objects. It is optimized for English, Simplified Chinese, and Traditional Chinese addresses but handles various formats. The agent determines whether two differently formatted addresses refer to the same physical location.

No. Customer messages, names, and addresses are sent to Rosette's API in real time during the conversation and used only for the current analysis. Tars does not maintain a separate store of the text content processed through Rosette. Results are returned to the conversation and not persisted.

Rosette uses deep neural networks and statistical models trained on multilingual name data. Accuracy varies by language pair but is designed for high-precision use cases like identity resolution, sanctions screening, and data deduplication. The similarity score gives you granular control over match sensitivity.

Yes. By enabling the multilingual parameter in the Language Identification request, Rosette detects language regions within a single document. The agent can identify that a message starts in English and switches to Spanish, allowing it to handle each segment appropriately.

Translation APIs convert text between languages. Rosette provides linguistic intelligence such as entity name resolution, address comparison, and language detection with confidence scoring. These are complementary capabilities. Your agent can use Rosette to identify the language and match names, then use a translation tool for response localization if needed.

Rosette uses API key authentication. You generate the key from your Rosette developer account and enter it in the Tars dashboard when adding the tool. The key is stored securely and used for all API calls. You can rotate or revoke the key from your Rosette account at any time.