OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Data Agents

DataFoundry

DataFoundry is an open-source, self-hosted AI data-analysis workbench that connects data sources with business definitions, tools and an agent runtime for interactive, multi-step analysis. It emphasizes read-only querying and auditable, replayable agent steps.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

DataFoundry — OpenViST project profile in Data Agents.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use DataFoundry when a team wants a self-hosted workspace for interactive data-agent analysis that incorporates enterprise definitions and keeps a reviewable record of the agent's work.

Who it is for

  • Data teams that want to analyze enterprise sources using their own metric definitions and schema context.
  • Organizations able to self-host an analytics workbench and configure a compatible model provider.
  • Teams that need to inspect or replay the SQL, tool calls and events behind an agent analysis.

Documented uses

  • Ask questions that require analysis across multiple tables and data sources.
  • Provide business definitions and field relationships to ground an agent's interpretation of metrics.
  • Review and replay analysis steps, then preview or export tables, charts, reports and SQL.
  • Continue or branch a prior analysis and refer to its output in a follow-up question.
Before you choose

Requirements & limits

The documented formal deployment path is self-hosted on Ubuntu or Debian. A model profile is created and enabled in the Web UI after deployment.

  • The README lists 28 datasource types, including PostgreSQL, MySQL, Snowflake, BigQuery, ClickHouse, MongoDB, Redis and Elasticsearch.
  • The project describes read-only queries, credential isolation, field masking, row limits and timeouts as default controls.
  • SQL, tool calls and event streams are retained for audit and replay.
  • DataFoundry is TypeScript-based, with Mastra identified as its agent runtime and AG-UI as its event-stream layer.
  • The documented deployment script builds the Web, API and TUI, but starts Web and API as background services while the TUI remains a separate foreground client.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Deploy DataFoundry on Ubuntu or Debian

Build and start the Web and API services using the documented formal deployment path.

Before you start

  • An Ubuntu or Debian environment.
  • Git and the repository's deployment script.

Steps

  1. Clone the repository: git clone https://github.com/datagallery-lab/datafoundry.git

  2. Change into the repository directory: cd datafoundry

  3. Run ./deploy.sh to generate configuration, install dependencies, build the Web, API and TUI, and start the Web and API as a detached process.

  4. Log in to the Web UI and create and enable a model profile.

Expected outcome

The Web and API are started, and a model profile can be configured in the Web UI for analysis.

Guide limitations

  • Docker and Compose are not provided in this release.
  • This documented deployment path does not support native Windows or macOS.
  • The TUI is built but runs as a separate foreground client; it does not remain running with the Web and API stack.
  • DataLink is external and is not started by the deployment script.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
datagallery-ai/dataagent · Official homepage title · retained in its source language
How we review sourcesSuggest a correction