Trieve AI search for the modern world

What is Trieve?

Trieve is a comprehensive AI-powered search and discovery platform that helps companies build competitive advantages through advanced search experiences. The platform combines billion-scale search capabilities with sophisticated features including semantic vector search, BM25 & SPLADE full-text search, and hybrid search functionality.

The platform stands out with its private open-source models approach, ensuring data security while providing powerful features such as sub-sentence highlighting, merchandising controls, and custom embedding model support. Trieve offers a modular infrastructure that can be self-hosted for organizations with sensitive data requirements or maximum performance needs.

Features

Semantic Vector Search: Advanced search capabilities beyond traditional full-text search
Hybrid Search: Combines full-text and semantic vector search with cross-encoder re-ranker models
Private Open-Source Models: Secure data handling with private server deployment
Merchandising Controls: API and no-code dashboard for search result optimization
Sub-sentence Highlighting: Precise result highlighting for improved user experience
Custom Embedding Models: Support for both stock and custom embedding models
Self-Hosting Options: Available for sensitive data and performance requirements

Use Cases

Enterprise search implementation
E-commerce product discovery
Document management systems
Knowledge base search optimization
Content recommendation systems
Large-scale data retrieval solutions

FAQs

What deployment options does Trieve offer?

Trieve offers both cloud-based deployment and self-hosting options through Docker Compose VPS, AWS EKS, and GCP GKE for organizations with sensitive data or specific performance requirements.
What is the data security approach?

Trieve uses open source embedding models and LLMs running on their own servers, ensuring that data is never leaked and maintaining complete privacy.
What is included in the free plan?

The free cloud plan includes 1,000 vectorized and stored chunks/records, 10 requests per second rate limit, 10 datasets, 5 users, 1,000 gen AI messages, and 1GB S3 file storage.