Documentation
Overview

Introduction to Altertable

Altertable is a AI-native data platform that combines storage, modeling, and analysis into an always-on system. Altertable is both an analytical database and a visualization layer, designed from day one to be used seamlessly by both humans and AI agents.

Altertable can be used for a wide range of analytical tasks, supporting use cases such as data analytics, business intelligence (BI), and product analytics.

The system is designed for continuous computation where data, models, and insights (charts, metrics, dashboards, etc.) stay live. To achieve this, Altertable runs a network of AI agents that continuously process and interpret data, keeping every dataset queryable, monitored, understood, and up to date.

Altertable is built on DuckDB for compute and Parquet for its state. DuckDB provides execution over this lake format using vectorized local queries and incremental snapshots. This allows Altertable to achieve warehouse-level performance with near-zero infrastructure overhead.

Agents rely on Anthropic, OpenAI, Google, and/or Mistral for their LLM capabilities.

Altertable's infrastructure has been designed and build from years of experience building the Distributed Search Network of Algolia, bringing the best of performance and scalability to the data platform space.

Learn the Core Concepts

New to Altertable? Start by understanding the core concepts that power the platform:

  • Insights: Understand the insight types and when to use each one
  • Agents: Learn how AI agents automate data work
  • Dashboards: Combine insights and monitor metrics
  • Discoveries: Review AI-generated findings
  • Architecture: Explore the lakehouse architecture and infrastructure

Getting Started

Choose your path based on your primary use case:

For Data Analytics & BI

For Product Analytics

For AI Integration

Crafted with <3 by former Algolia × Front × Sorare builders© 2026 AltertableTermsPrivacySecurityCookies