Trieve

Trieve, located in San Francisco, CA, operates within the B2B infrastructure industry, specifically targeting search solutions. As a participant in Y Combinator’s W24 batch, Trieve has established itself in both the United States and Canada. The company, though small with a team of three, offers robust search technologies including SPLADE full-text neural search and semantic vector search, which can be combined in a hybrid search model. Trieve’s services are designed to enhance search capabilities through features like merchandising relevance tuning, date recency biasing, and recommendations based on user history and content similarity. They also provide a novel approach to duplicate detection and merging, and support sub-sentence highlighting in search results. Trieve emphasizes data privacy by hosting open source embedding models and large language models on their own servers. They offer a comprehensive API, extensive documentation, and support for self-hosting with terraform templates without external dependencies. Additionally, Trieve provides a free tier, various pricing plans, and the ability to book demos directly through their website, making it accessible for businesses to start integrating their services. The company also maintains a Github repository for source code access and contributions, ensuring transparency and community engagement.