Search Toolkit Released

Search Toolkit, a composable framework for building production-grade search pipelines for AI applications, has been released in public preview.

When building traditional search infrastructure, developers had to individually integrate separate tools for data collection (Ingestion), search (Retrieval), and evaluation (Evaluation), consuming enormous engineering resources in the process. Search Toolkit unifies these three stages into a single framework with a shared interface, allowing developers to focus on improving search quality instead of maintaining infrastructure.

Key features and benefits include:

• Enterprise Search: Provides consistent processing and indexing patterns across diverse sources such as wikis, codebases, and document repositories, reducing the burden of rebuilding pipelines.
• RAG and Search Quality Management: Built-in evaluation capabilities allow search performance to be measured independently, making it clear whether issues arise from Retrieval or Generation.
• Domain-Specific Search: Supports effective search in specialized fields such as legal, medical, and finance that require domain-specific terminology and unique document structures.

Search Toolkit is available as open source and can run anywhere on the user's infrastructure, including cloud, on-premises, and edge environments.