Redpanda SQL
Product · SaaS / cloud software · Data & analytics by Redpanda
The query engine that skips the pipeline for real-time AI and analytics.
A PostgreSQL-compatible distributed OLAP engine for real-time analytics on live and historical data. Built for AI agents and data teams to skip ETL pipelines and query streaming data with zero lag.
- Launched
- 2026-05-27
- Languages
- English
Category: Database
- Deployment
- Cloud (Serverless, Dedicated, BYOC), Self-Managed
- Platforms
- API, Web (Redpanda Console)
- APIs and SDKs
- PostgreSQL-compatible interface, Kafka API-compatible engine
- Editions
- Generally Available (GA) as part of Redpanda Data Platform
- Engine Type
- Distributed OLAP
- Compatibility
- PostgreSQL-compatible
- PostgreSQL-compatible distributed OLAP engine
- Queries live topics and historical Iceberg tables together
- Zero-ETL architecture eliminates pipeline lag
- Supports parallel SQL fan-out for AI agents
Overview
Redpanda SQL is a PostgreSQL-compatible distributed OLAP engine designed for real-time analytics on streaming data. It serves businesses needing to query live topics and historical data in Apache Iceberg format without the latency of traditional ETL pipelines. By providing a single interface for both streaming and historical data, it ensures that analytics and AI agents operate on the most current state of the business. The engine is particularly optimized for agentic workloads, enabling AI agents to fan out across hundreds of parallel queries to retrieve fresh context for RAG pipelines and automated decision-making. It is delivered as a core component of the Redpanda Data Platform, available across major cloud providers and self-managed environments.
Real-Time Analytics Engine
- PostgreSQL Compatibility: The PostgreSQL-compatible distributed OLAP engine allows users to query live topics and historical data using familiar SQL syntax.
- Live and Historical Queries: Users can perform ad-hoc SQL queries against live streaming topics and historical data stored in Apache Iceberg tables without union logic.
- Zero-ETL Analytics: The engine eliminates the need for complex ETL pipelines by allowing users to query data directly where it lives, skipping traditional pipeline lag.
- Bridge Queries support spanning multiple data sources for comprehensive real-time analysis across the data platform.
AI and Agent Integration
- Agentic SQL Support: AI agents can fan out across hundreds of parallel SQL queries to explore hypotheses and retrieve fresh context from live data.
- AI Observability: Every agent execution and query is traceable to its source data in the event log for built-in AI observability.
- Agent Control: The engine provides a model-based kill switch capability to freeze agents when they break rules or exceed budgets.
Use cases
- AI agents perform parallel SQL queries across live streams to retrieve fresh context for RAG pipelines.
- Data analysts run ad-hoc queries against live topics and historical Iceberg tables without building separate ETL pipelines.
Specifications
- Deployment
- Cloud (Serverless, Dedicated, BYOC), Self-Managed
- Platforms
- API, Web (Redpanda Console)
- Integrations
- Apache Iceberg, PostgreSQL, Snowflake, Salesforce, Google Drive, MongoDB, MySQL, AWS, GCP, Azure
- APIs and SDKs
- PostgreSQL-compatible interface, Kafka API-compatible engine
- Editions
- Generally Available (GA) as part of Redpanda Data Platform
- Engine Type
- Distributed OLAP
- Compatibility
- PostgreSQL-compatible
- Data Formats
- Live topics, Apache Iceberg tables
- Architecture
- Zero-ETL, single-engine duality
- Query Interface
- SQL
How it works
- Deployment: Deploy Redpanda SQL as part of the Redpanda Data Platform on your preferred cloud (AWS, GCP, Azure) or self-managed infrastructure.
- Data Integration: Connect your live data streams and historical data lakes using native Apache Iceberg integration.
- Querying: Use standard PostgreSQL-compatible tools or AI agents to run ad-hoc queries against live and historical data simultaneously.
FAQ
How does it handle historical data?
It allows users to query live streaming data and historical Apache Iceberg tables using a single SQL interface, eliminating the need for separate ETL pipelines.
Is it compatible with AI agents?
Yes, it is specifically optimized for agentic workloads, allowing AI agents to perform parallel queries to get fresh context without waiting for warehouse lag.
How is it deployed?
Redpanda SQL is generally available as part of the Redpanda Data Platform, which can be deployed as Serverless, BYOC, or Self-Managed.
Added by AntSpace from public sources (redpanda.com).