DOCSEARCHbyAlgolia
Submit Docs

Evidence Documentation

website·Indexed 35 days ago

https://docs.evidence.dev

Documentation for Evidence, a tool for building data reports. Includes guides on the Evidence CLI for local development and warehouse connectivity, as well as a comprehensive library of components for data visualization, including area, bar, bubble, and combo charts, heatmaps, and interactive filters.

Tokens
131.5K
Snippets
597
Records
1.2K
Agent score
67%

What's inside Evidence

  1. Overview of Evidence Warehouse

    The Evidence Warehouse is a built-in, fully managed cloud warehouse available exclusively for Evidence Studio users. It handles the ETL (extract, load, transform) process by extracting data from connected sources and storing it in a chosen region. Reports query this managed warehouse for high performance, removing the need for users to manage their own infrastructure for data integration.

  2. Overview of Evidence Embedded Analytics

    Embedded Analytics allows you to deliver personalized Evidence reports to customers directly within your own application using an iframe embed. It supports white-labeling through custom themes and organization logos. The system uses a two-step authentication flow: a server-side request to the Evidence Embed API to generate a secure URL, followed by a client-side load of that URL in an iframe. Key features include Row Level Security (RLS) to filter data by user, support for multiple languages, and JWE encryption for user attributes.

  3. Overview of Studio MCP for AI Assistants

    The Studio MCP allows AI assistants like Claude and ChatGPT to connect to a live Evidence Studio workspace. Once connected, the assistant can access projects, read published reports, and query data. The server is hosted at https://evidence.studio/api/mcp and uses Streamable HTTP. All requests are executed as the signed-in user, meaning existing access policies and page controls are respected.

  4. Overview of Scheduled Reports in Evidence Studio

    Scheduled Reports allow you to deliver published reports via email or Slack on a recurring schedule (daily, weekly, monthly) or as a one-off copy. This feature is available exclusively in Evidence Studio (the hosted platform). Reports are rendered as PDFs for delivery. Key capabilities include:

    • Delivery to specific users, groups, or Slack channels.
    • Support for shared renders (one view for all) or per-recipient renders (personalized data via row-level security).
    • Viewer self-subscriptions for non-admin users.
    • Run history tracking with status and delivery results.

  5. Overview of the Evidence Agent

    The Evidence Agent is a business intelligence AI agent available in Evidence Studio. It allows readers to ask data questions in natural language and receive analytical answers (insights) written in Evidence Markdown. The agent can access underlying SQL and filter definitions while respecting the project's existing row-level security and page-level access control. Its behavior is customized via a version-controlled agent/ directory at the project root.

  6. Overview of Evidence Models

    Models are SQL queries that run in the background and materialize results, allowing developers to prepare optimized, reusable datasets. Once created, models can be referenced in visualizations, the SQL console, Explore, and reports just like regular data sources. They are particularly useful for joining sources, type casting, centralizing business logic, and pre-running heavy transformations to avoid slowing down page loads.

  7. Overview of Page Level Access Control in Evidence Studio

    Page level access control in Evidence Studio uses a Role-Based Access Control (RBAC) as code model. Permissions are defined in a single access.yaml file at the project root. This file is version-controlled via git, allowing access changes to be reviewed in pull requests.

    Key Access Rules:

    • Users with Developer, Admin, or Org Viewer roles always have access to all pages, regardless of the access.yaml configuration.
    • Internal viewers are employees/teammates within the organization.
    • External viewers (Customers) are clients or partners. They are excluded from any $org (everyone in the org) rules and must be granted access explicitly.

  8. Use Table Actions for data export and viewing

    Table footers now include 'Download' and 'Fullscreen' buttons to allow users to export data or view the table in a larger format.

  9. Use Markdown and Markdoc components in Evidence

    Evidence pages are written in Markdown and support almost all standard Markdown syntax (lists, bold, italics, inline code, and links). For visual elements and charts, Evidence uses a built-in component library based on Markdoc syntax. Components can be added manually or via slash commands in the editor.
    {% line_chart
    data="demo.daily_orders"
    x="date"
    y="sum(sales)"
    date_grain="month"
%}

  10. Connect YouTube Analytics data to Evidence

    The YouTube Analytics Connector allows you to sync channel reports, content owner reports, and metadata tables into Evidence.

  11. Connect to Parquet files using the Custom Bucket Provider

    The Custom Bucket Provider allows Evidence to read Parquet files directly from cloud storage. It utilizes ClickHouse's S3 table engine to query data in place, meaning no data is copied into Evidence.

  12. Use Evidence Studio MCP Server for AI clients

    The Evidence Studio MCP Server allows you to search reports and query data from AI clients like Claude and ChatGPT. Every request is executed as the signed-in user, ensuring that access policies and page controls are respected.