DocumentationArchitectureNews & ResearchFAQAbout
Hotdata
Getting Started›Introduction

Introduction

Hotdata gives developers and agents a single API to access data across databases, warehouses, and SaaS applications as if it were one system. You can join tables from different sources in a single query without moving data or maintaining pipelines.

Hotdata connects directly to your sources. There’s no need for ETL, replicas, or separate data infrastructure. Data is ingested on demand when first queried, or proactively cached for selected tables. From there, it stays up to date based on schedules or observed usage patterns.

This allows you to join data across systems—databases, warehouses, and files—through one interface, while avoiding the operational overhead of traditional data movement. Each workspace runs in its own isolated execution environment with sub-second query latency.

  • Connect to Postgres, BigQuery, Snowflake, and other systems
  • Query across sources using a single interface
  • Persist datasets for reuse, audit, and replay

A workspace is the core unit of isolation:

  • Create a workspace
  • Attach connections
  • Inspect tables
  • Run queries
  • Save queries as reusable datasets
  • Manage secrets
  • Delete the workspace

Why use Hotdata?

  • Sub-second latency. Each workspace runs as an isolated runtime optimized for fast responses. This matters when queries sit inside an application or agent loop.
  • Designed for short-lived workloads. Agents and apps issue many concurrent, temporary queries. Hotdata scales with that pattern without the overhead of batch-oriented systems.
  • Usage-based model. No always-on infrastructure. Workspaces are created when needed and removed when no longer in use.
  • Isolation by default. Every workspace acts as a security boundary. Requests are scoped to that environment, giving agents controlled, repeatable access without shared state.

What you can do today

Hotdata supports a web API, MCP, SQL, and CLI.

Examples:

  • List workspaces: hotdata workspace list
  • List connections: hotdata connections list <workspace_id>
  • List tables: hotdata tables list <workspace_id>
  • Execute SQL: hotdata query "<sql>" --workspace-id <workspace_id>

See Quick Start to get started, CLI Reference for commands, and API Reference for HTTP endpoints.