DataAgent v0.1.0

The first official release of DataAgent, establishing a versioned baseline for building configurable Agents across NL2SQL, semantic data understanding, and end-to-end data workflows.

Introduction

DataAgent has been developed in the open as a framework for building data-oriented AI agents.

DataAgent brings together NL2SQL, semantic data understanding, and configurable Agent capabilities to turn natural-language requests into end-to-end data workflows. It helps teams connect models, data sources, and domain knowledge to build reusable Agents for enterprise data Q&A, data analysis, and scientific data exploration.

Release Highlights

Configurable Agents

DataAgent lets users define and run data agents through YAML configuration. Models, tools, memory, workspace boundaries, and scenario instructions can be adapted to different data tasks without changing the Agent's core code.

Data and NL2SQL Capabilities

This release includes a dedicated NL2SQL pipeline for:

• Schema perception
• SQL generation
• SQL validation and execution
• Result explanation

DataAgent can optionally connect to the DataGallery Semantic Service for enriched metadata, semantic schema retrieval, relationship discovery, and SQL few-shot recall.

The Semantic Service is optional for general Agent workflows but required for semantic-layer-assisted NL2SQL scenarios.

Context and Trace Management

The Context system manages conversation state, business metadata, tool information, and execution traces.

It maintains structured DAG and IR representations for traceability and supports external persistence and checkpoint backends for stateful workflows.

Multiple Ways to Build and Run

DataAgent v0.1.0 provides:

• Interactive CLI
• YAML-based Agent configuration
• Python SDK
• Configuration validation
• Streaming and non-streaming invocation
• A2A service mode

Release Scope

DataAgent v0.1.0 establishes the initial release baseline. At this stage:

• Python 3.11 or later is required.
• A compatible model endpoint and API key are required.
• The Semantic Service is deployed separately and is optional for general Agent workflows.

Future releases will document new features, bug fixes, compatibility changes, and migration requirements relative to this baseline.

Documentation

• Installation Guide
• Quick Start
• Core Features
• Architecture
• Python SDK and A2A APIs
• Build an NL2SQL Application

Chinese and English documentation are available under the docs/ directory.

Contributors

Thank you to everyone who contributed code, tests, documentation, examples, reviews, and feedback leading up to the first official DataAgent release.