OpenSearch Veterans Launch Infino, an Agent Retrieval Platform for AI Agents

Infino Launches Unified Retrieval Platform

Infino, a company founded by OpenSearch veterans, officially launched its agent retrieval platform. The platform aims to provide a single retrieval layer, functioning as a unified database system that allows all AI agents to query data from one central location.

Addressing Fragmented Data Stacks for AI Agents

According to Infino CEO Ekechi Nwokah, AI agents represent the "largest new consumer of data since the web browser." However, these agents are currently served by a fragmented data stack, including data warehouses for SQL, search engines for keywords, and vector databases for semantic search, all requiring ETL pipelines for synchronization. Infino's platform seeks to collapse this fragmented stack.

Optimized for Agent Query Patterns

Traditional data stores are optimized for human or application queries that involve a single request and result. Nwokah explains that AI agents operate differently, issuing many small, concurrent questions to answer a larger one. Each result consumes context, potentially adding cost and latency to subsequent model calls. Infino's retrieval layer is designed specifically for this dynamic, allowing agents to search, rank, filter, join, aggregate, and reason over data stored in Parquet in object storage directly, using a single copy of the data.