Visual Diff API, Automation Rename & Audit Fixes
Feature
Fix
Improvement
- ✓Visual Diff API — New
POST /api/v1/diffendpoint. Send two URLs (or HTML strings) and get a pixel-level diff image with changed pixels highlighted in red, pluschanged_pct,changed_pixels, andtotal_pixelsmetrics. Supports all standard options (viewport, dark mode, banner blocking, cookies, etc.). Costs 1 request. - ✓Visual Diff fix — Fixed a bug where the Visual Diff tool failed silently. Puppeteer 23 returns
Uint8Arrayfrompage.screenshot(), but pngjs requires a Node.jsBuffer. Now wraps the result withBuffer.from(). - ✓Dashboard: “Automation” rename — Renamed “Sequence & Video” to “Automation” across the dashboard. The toggle is now framed as an output mode choice: screenshots/PDFs or recorded video.
- ✓SDK visibility — Added official SDK callouts to the homepage and dashboard. Node.js (
npm install pagebolt) and Python (pip install pagebolt) install hints with docs links. - ✓MCP server v1.8.2 — Fixed
list_devicesproperty paths,list_sessionsJSON parsing, PDF blob returns in sequences, and standardized error handling. Added request timeouts (120s) and retry logic for 429/5xx. - ✓Pricing accuracy — Corrected PAYG pricing from flat “$0.01/req” to per-endpoint breakdown. Updated Growth API key limit from 5 to 10, Scale from 5 to 25. Fixed “6 tools” → “7 APIs” on the Try page.
- ✓Trial safety net fix — Fixed undefined
db.endTrialin the trial expiry safety net. Cookie banner blocking now coverscustodia-privacy.com.