Evidence is business intelligence as code: your entire reporting platform — pages, queries, access control and themes — defined in a Git repository as Markdown, SQL and components.
Drag-and-drop BI tools create dashboards nobody can review, version or test. Evidence takes the opposite approach. Analysts author reports as Markdown files with embedded SQL and a component library, and the whole project lives in a repo with pull requests, CI checks, branch previews and instant rollbacks — the same workflow engineers already trust. Push to deploy, and merging a change ships it to readers immediately.
The current platform adds an analytics agent on top of that foundation. The agent understands your project — every existing dashboard and report becomes its context — and answers questions or builds new reports for the whole team, reachable from Slack or any MCP client such as Claude Desktop and ChatGPT. Teams can extend it with custom context and reusable ‘skills’: analytical playbooks with visualizations, inputs and clarifying questions. Around the core sit publication-grade visualizations, rich tables, maps, scheduled email delivery, PDF/XLSX exports, page-level access control down to the row, and embedding or white-labeling for customer-facing analytics, backed by enterprise SSO, SCIM, audit logs and SOC 2.
Evidence Core, the open-source framework, is free to self-host. The commercial platform — cloud Studio editor, hosted deployment, the analytics agent and enterprise security — is priced per deployment; third-party pricing databases list commercial plans starting around $500/month, with enterprise agreements quoted individually. Teams typically evaluate through a demo before committing.
Data teams who want BI under version control and agents that can safely build on it, plus SaaS companies embedding analytics for customers. Teams wanting a purely visual, no-code dashboard builder will find Evidence’s code-first approach a deliberate mismatch.