Skip to content

Evidence

Write data reports using SQL and Markdown — results update automatically from live data sources.

Overview

Evidence turns SQL queries and Markdown prose into polished, interactive data reports. Write a report template once — embed your SQL queries inline, add charts and tables, and the results populate automatically from your connected data sources. When your data changes, your report updates. It’s the fastest way to build data narratives that stay current.

Key Features

  • SQL + Markdown — Mix query results directly into narrative text
  • Live Data Refresh — Reports pull fresh data on every build or page load
  • Interactive Charts — Built-in chart components (line, bar, scatter, area, and more)
  • Multi-Source Queries — Query multiple databases in a single report
  • Static Export — Publish reports as fast, static HTML sites
  • Version Control Friendly — Reports are plain text files you can commit to git

Getting Started

  1. From the Hub, click Evidence to launch
  2. Evidence opens with your report workspace
  3. Open or create a .md file in pages/
  4. Write a SQL query block to pull data
  5. Add chart components to visualize it
  6. Preview your report live in the browser

Writing a Report

Evidence reports are Markdown files with embedded SQL:

# Monthly Revenue Report

```sql orders
select
  date_trunc('month', created_at) as month,
  sum(total) as revenue
from orders
where created_at >= '2024-01-01'
group by 1
order by 1
```

Revenue has grown consistently over the past year.

<LineChart data={orders} x="month" y="revenue" />

The SQL query runs against your configured data source and the result is available as a variable (orders above) that you pass to chart components.

Connecting Data Sources

Add your database connections in evidence.config.yaml:

sources:
  - name: warehouse
    type: postgres
    host: your-db-host
    database: analytics

Supported sources include PostgreSQL, MySQL, BigQuery, Snowflake, DuckDB, SQLite, CSV files, and more.

Chart Components

Evidence includes a full component library:

  • <LineChart> — Time series and trend lines
  • <BarChart> — Comparisons and rankings
  • <ScatterPlot> — Correlations
  • <DataTable> — Sortable, filterable tables
  • <Value> — Inline metric display
  • <BigValue> — KPI callout cards

When to Use Evidence

TaskTool
Recurring data reports with narrativeEvidence
BI dashboards with interactivitySuperset
Ad-hoc SQL explorationDB Loadr
Notebook-based analysisAI Notebook Lab
Sharing with non-technical stakeholdersEvidence or Superset