Skip to content

Periscope MCP Server ★ New (July 2 PM)

Overview

Periscope MCP (segentic-lab/periscope-mcp) delivers 63 Playwright-powered website testing tools designed specifically for AI agents. Unlike general browser automation MCPs, Periscope is built with agent-first ergonomics — each tool is a discrete testing action (navigate, assert, screenshot, check accessibility, measure performance) that composes cleanly into AI-driven test workflows. Development and QA teams can write, run, and debug end-to-end tests through natural language, with results structured for AI consumption.

Key advantage: 63 purpose-built testing tools, agent-first design, Playwright-backed reliability — testing that works the way AI agents think.

Key Features

  • Browser navigation: Navigate to URLs, click elements, fill forms, scroll, wait for conditions
  • Screenshots & visual regression: Capture full-page and element screenshots; compare against baselines for visual regression detection
  • Assertions: Assert element visibility, text content, attributes, CSS properties, and page state
  • Accessibility checks: Run axe-core accessibility audits, flag WCAG violations
  • Performance metrics: Capture Core Web Vitals (LCP, CLS, INP), page load timings, and resource waterfall
  • Network monitoring: Intercept and inspect network requests, mock API responses, validate API calls
  • CI/CD ready: Structured JSON output for test runners, GitHub Actions, and CI pipelines
  • Agent-first ergonomics: Each tool returns structured, parseable results — no parsing raw HTML or console output

Installation

# Requires Playwright browsers
npx playwright install

# Clone and install
git clone https://github.com/segentic-lab/periscope-mcp.git
cd periscope-mcp
npm install
npm run build

# Add to Hermes
hermes mcp add periscope --command "node" --args "dist/index.js" --workdir "$(pwd)"

Configuration

{
  "mcpServers": {
    "periscope": {
      "command": "node",
      "args": ["dist/index.js"],
      "workdir": "/path/to/periscope-mcp",
      "env": {
        "BASE_URL": "https://staging.example.com",
        "HEADLESS": "true",
        "SCREENSHOT_DIR": "./screenshots"
      }
    }
  }
}

Business Relevance

  • QA Automation: QA teams can write and execute test suites through natural language — "test the checkout flow on mobile viewport and flag any layout issues" — without writing Playwright scripts
  • Visual Regression: Automatically detect visual regressions after deployments — screenshot comparison with pixel-diff reporting
  • Accessibility Compliance: Run WCAG audits as part of CI/CD — catch accessibility regressions before they reach production
  • Pre-release Validation: AI agents can run a full test battery before every release — login flow, critical paths, form validation, mobile responsiveness
  • Competitive Monitoring: Schedule automated checks on competitor websites — feature changes, pricing updates, UX modifications

Integration with CorpusIQ

Periscope MCP strengthens CorpusIQ's development and QA toolchain, complementing existing testing infrastructure:

  • CI/CD Pipeline: Combine Periscope's browser testing with Drumbeats MCP for monitoring — "if deployment succeeds, run Periscope smoke tests, then update Drumbeats status page"
  • Pre-flight Testing: Integrate with CorpusIQ's pre-flight gate — before publishing content or deploying changes, run Periscope checks on the affected pages
  • Competitive Intelligence: Pair with SEO tools (SE Ranking, OpenSEO) — Periscope captures visual/functional competitor changes while SEO tools track ranking shifts
  • Agent Workflows: AI agents can orchestrate end-to-end validation — database query (dbridge) → API check (Bruno) → browser test (Periscope) → deploy (Cloudflare)

Limitations

  • Requires Playwright browser binaries (Chromium, Firefox, WebKit) — ~500MB disk space
  • Self-hosted — runs locally, not a cloud service
  • Chromium-based rendering may differ from real device browsers
  • Visual regression sensitive to anti-aliasing and font rendering differences across OS

See Also