What it is

Evidence is an open-source, code-based business intelligence tool for building reports and data visualizations. It lives in the Data & Analytics ecosystem, specifically Business Intelligence and Reporting, and is written in TypeScript under the MIT license. Reports are authored in SQL and markdown rather than assembled through drag-and-drop interfaces, and the project describes itself as an agent-ready tool for creating reports from data. The repository carries 6,936 stars and 417 forks, with 37 open issues at the time of writing.

The concrete problem it solves is the gap between version-controlled, reviewable analytics work and the point-and-click BI products that dominate the category. Teams that already keep SQL, dbt models, and documentation in git can build interactive dashboards in the same medium, so a report becomes a file that can be diffed, reviewed, and deployed like any other artifact. Evidence positions itself explicitly as an open-source alternative to drag-and-drop BI tools, which means the workflow stays in the editor and the terminal instead of a proprietary web builder. The topic list points at the surrounding ecosystem it plugs into: dbt, DuckDB, exploratory data analysis, data engineering, and Svelte on the rendering side.

Key capabilities

• Author reports in SQL and markdown, keeping queries and narrative in the same source files.
• Build fast, interactive data visualizations as described in the project description.
• Develop with the Evidence Agent, or locally with tools such as Claude Code and Cursor.
• Publish to Evidence Studio for hosted reports, or self-host the generated static site.
• Install and manage projects through a dedicated evidence command-line interface.
• Connect to the dbt and DuckDB tooling listed among the project topics.
• Render output through a Svelte-based front end, per the repository topics.

Who uses it and how

• Analytics and data engineering teams that keep SQL and dbt models in version control and want reports to live alongside them.
• Data scientists and analysts doing exploratory data analysis who need shareable, interactive output rather than static notebooks.
• Developers who prefer editing reports in an IDE or an AI coding assistant instead of a browser-based dashboard builder.
• Organizations that need to self-host their reporting stack on their own infrastructure.
• Teams that want a hosted option and choose Evidence Studio for publishing instead of running their own deployment.

Getting started

On macOS and Linux, install with curl -fsSL https://evidence.studio/install.sh | sh, and on Windows with irm https://evidence.studio/install.ps1 | iex. Then run evidence init my-project, cd my-project, and evidence dev to start developing; publishing is done either on Evidence Studio or by self-hosting the generated static site.

When to use it — and when not to

Evidence is a reasonable fit when the team already writes SQL and wants an open-source replacement for drag-and-drop BI products, and when reports need to be reviewed and versioned like code. A self-hoster must operate the generated static site on their own infrastructure, and the publishing documentation should be consulted for the available options. The repository shows 37 open issues and a recorded contributor count of zero, so anyone evaluating it for production should check current activity and release cadence rather

curl -fsSL https://evidence.studio/install.sh | sh
evidence help

irm https://evidence.studio/install.ps1 | iex
evidence help

evidence init my-project
cd my-project
evidence dev