hotdata

Hotdata provides an API platform designed to enable agents and applications to create databases on demand, supporting high-volume, low-latency search and query operations across diverse data sources. Its core offering addresses the challenge of managing concurrent queries from agents that need rapid, consistent access to data without the overhead of custom integration or infrastructure provisioning. The platform is positioned for use cases where agents require fast, read-only exploration and analysis of data from multiple systems.

The service allows agents to programmatically create instant databases, declare tables, and load files such as Parquet for immediate querying. These databases are temporary and scoped to the agent, promoting safe experimentation without risk to production systems. Hotdata supports advanced query capabilities, including OLAP, vector search, top-k, and geospatial operations, which can be combined within a single query. The platform operates in read-only mode, providing access to snapshots of connected databases, warehouses, and SaaS systems, ensuring that agents cannot modify or delete source data.

Hotdata can be used as a fully managed service or deployed within a user's own cloud environment, offering flexibility in infrastructure control and data governance. Its architecture is built on Rust, Apache Arrow, and Apache DataFusion, with Arrow-native execution underpinning its query engine. The platform integrates with existing data stacks by layering on top of current databases and applications, acting as a unified, low-latency query interface without replacing existing systems of record.

10 per 10 GB scanned per month. A base platform fee is listed, and there is mention of a sign-up credit that can offset initial costs. Hotdata is currently available in early access. The tool is intended for developers, teams, and AI agents that require scalable, secure, and efficient access to data for analytical and exploratory purposes.