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.
