AI-powered Search & RAG Platform
2.6k 2026-04-30
devflowinc/trieve
An all-in-one API platform for building intelligent search, recommendations, and RAG applications with advanced semantic capabilities.
Core Features
Self-hosting options for AWS, GCP, Kubernetes, and Docker Compose.
Hybrid search combining semantic vector search (Qdrant) and neural sparse-vector search (SPLADE).
Convenient RAG API routes with LLM integration (OpenRouter) and topic-based memory management.
Recommendations API for finding similar content and sub-sentence highlighting for enhanced UX.
Tunable relevance using signals, recency biasing, and comprehensive filtering/grouping capabilities.
Detailed Introduction
Trieve is an API-first platform designed to empower developers with advanced search, recommendation, and Retrieval Augmented Generation (RAG) functionalities. It offers a comprehensive suite of tools including semantic vector search, typo-tolerant full-text search, and hybrid search with re-ranking. With flexible self-hosting options, integration with various embedding and LLM models, and features like recency biasing and content grouping, Trieve provides a robust foundation for building highly intelligent and customizable data interaction experiences across diverse applications.