Conversation Data Extraction
Extracts named entities, topics and sentiment from chat messages, pseudonymizes personal data (GDPR) and stores structured results.
Anonymization happens locally, just like with the local models, before the pseudonymized text is sent through the API. The chosen model then determines sentiment, entities and topic in one call.
Sentiment model
NER (entity extraction) model
Topic classification model
Engine
Sentiment
Topic
Entities
Anonymized text
Batch ingest
Upload a CSV or JSON file of conversations. Each message is processed by
the full pipeline and stored (anonymized). CSV should have a
text or message column.
0.0 s
Processed 0 of 0