Live · Visual QA and browser tooling
Pursr
A visual QA toolkit for browser capture, accessibility audits, screenshot diffing, persistent browser sessions, and agent-driven inspection through MCP.
RepositoryWhy it exists
Browser review often requires separate tools for screenshots, visual regression, accessibility checks, interaction diagnostics, and reports.
Pursr brings those workflows into one CLI, library, and MCP server so developers and coding agents can inspect the same rendered evidence.
Current scope
The project drives a locally installed Chromium-family browser through Playwright and keeps browser installation under the operator’s control.
- Multi-viewport screenshots and visual baselines
- Pixel diffs and batched visual sweeps
- Accessibility audits with highlighted evidence
- Persistent MCP browser sessions and direct image responses
- HTML, CI, and PDF reporting
How it is used
A developer can use the CLI for repeatable capture plans or attach an agent through MCP for inspect, act, screenshot, diagnose, and verify loops.
Status
Pursr is published as a public Node.js package and remains under active development around visual operator workflows, browser diagnostics, and reusable QA evidence.