Trieve

What is Trieve?

Trieve offers a powerful, AI-first infrastructure API that seamlessly handles search, recommendations, and retrieval-augmented generation (RAG) to enhance discovery experiences for companies. It combines advanced language models with tools for fine-tuning ranking and relevance. The platform supports semantic vector search and hybrid search, integrating full-text search with semantic capabilities through cross-encoder re-ranker models. Users can tune and boost search results to meet key performance indicators using either the API or an intuitive no-code dashboard. The system is designed for easy management of data ingestion, embeddings, and analytics. Employing open-source models run on private servers ensures data security, while providing options for self-hosting and maximum performance, making adaption swift and easy.

Key features

• Billion-scale search and RAG within the API.
• Supports semantic vector search out of the box.
• Choose from BM25 & SPLADE full-text search.
• Hybrid search with cross-encoder re-ranker models.
• Tune and boost results with a no-code dashboard.
• Sub-sentence highlights for quick results navigation.