The query engine for
Agentic Workloads
Lookups, OLAP, full-text, and vector search in a single Postgres query
Let your agents cook - 20× faster, at 90% lower cost.
Install with Homebrew
brew install hotdata-dev/tap/cli
20×
Faster
than federated engines
90%
Lower Cost
than data warehouses
60+
Connectors
to databases, warehouses & SaaS
Features
Secure Sandbox
Agents run in an isolated sandbox to protect production systems, with a complete audit trail of every action.
Unified Query API
A single interface for lookups, full-text search, OLAP, geospatial, and vector search—across databases, data lakes, warehouses, and SaaS.
Millisecond Responses
Queries are cached and instantly reusable with no redundant computation or waiting.
Architecture Diagram
Fully managed or deployed in your own cloud environment.
Built with Rust, Apache Arrow, and Apache DataFusion
Questions?
We've found that agents tend to query many sources at once and generate large numbers of concurrent requests that need fast responses. Teams trying to support that pattern often end up building a lot of custom infrastructure—connectors, pipelines, caches, and routing logic. We built Hotdata to simplify that, providing a single API that can return low-latency results across many data sources.
Our query engine is built on Apache Arrow. It allows us to execute queries efficiently while pulling data from different systems on demand.
We've seen agents run many short queries while exploring data and generating intermediate results. Supporting that pattern requires fast response times and the ability to access different data sources without building separate pipelines. Hotdata provides a single execution layer that allows agents to query those systems quickly and consistently.
We've designed Hotdata so agents operate in isolated environments with read-only access to underlying data sources. Agents cannot modify your systems of record. They can create and discard temporary datasets as part of their workflow, but those datasets are scoped to the agent and can be safely destroyed without affecting source data. This allows agents to explore, transform, and iterate on data while keeping the underlying systems protected.
No. We focus on fast read access. Hotdata operates in read-only mode and takes snapshots of your databases, warehouses, and SaaS systems. Updates, deletes, and writes continue to happen in your system of record.
No. We only provide read access to connected data sources.
We've designed Hotdata to sit on top of your existing systems of record and access data on demand. Your databases, warehouses, and SaaS systems remain the source of truth while Hotdata provides a low-latency execution layer for applications and agents.
Under the hood we use Apache DataFusion with Arrow-native execution.
Yes. With BYOC we can deploy Hotdata directly into your cloud environment while we manage the service. You retain control over infrastructure, access, and keys. We also offer a fully managed option.