Trieve

Trieve is infrastructure for adding AI-powered search and chat experiences to digital products. It combines semantic search, conversational answers, and recommendations so users can find content or products faster.

Search and chat on one AI core

Trieve uses modern language models to understand natural-language queries. User requests are converted into vector representations, which helps return relevant results even when wording is imprecise. The same core can power chatbots that answer using your company data and keep context across a conversation.

Recommendations and product analytics

You can connect catalogs, documentation, knowledge bases, and user behavior logs. Based on this data, Trieve generates dynamic recommendations and improves search relevance over time. Built-in analytics helps teams see what users search for, where results fall short, and which changes improve key product metrics.

Open source and project status

Trieve is being shut down as a cloud service and is moving to open source under the MIT license after being acquired by Mintlify. Teams can self-host Trieve, adapt it to their infrastructure, and use it as a foundation for custom search and chat solutions.