HeadlessX v2.1.1

Released: 2026-03-20

Highlights

HeadlessX v2.1.1 focuses on operator-first API cleanup, Google AI Search improvements, a published CLI flow, Docker completeness, and runtime stability across the dashboard and backend.

What Changed

Operator-first routing

• reorganized public API routes under /api/operators/*
• added operator status at /api/operators/status
• reorganized playground routes under /playground/operators/*
• stabilized Website Scrape as a single frontend route with format switching inside the UI

Persistent browser runtime

• added one shared persistent Camoufox profile in the API
• kept cookies, storage, and browser state on disk
• made runtime prefer apps/api/data first, then apps/api/default-data
• removed the old forced 1920x1080 browser size and improved viewport behavior

Google AI Search

• renamed the old Google flow to Google AI Search
• added region, language, and time-filter targeting
• kept plain google.com as the default start flow unless targeting is explicitly set
• improved SSE stability and AI overview rendering
• made CAPTCHA progress messaging appear only when a real CAPTCHA is detected
• aligned the Google workbench UI with the Website UI

CLI

• published the CLI as @headlessx-cli/core
• standardized the command as headlessx
• made CLI output markdown-first by default
• updated login, auth, operator, and Google AI Search command flows

Crawl and queue reliability

• fixed crawl cancel behavior
• fixed the enqueue/cancel race window
• improved worker-side cancellation handling
• fixed crawl page budgeting
• made Website Crawl use the advanced timeout value correctly

Docker and self-hosting

• added yt-engine to the full Docker stack
• made full Docker closer to one-command setup
• aligned setup and API docs with the current runtime model
• documented that Website Crawl specifically requires Redis and the worker

Settings and UX cleanup

• simplified proxy settings UI
• replaced browser-native alerts with in-app feedback
• fixed operator Stop button behavior across workbenches
• cleaned settings layout and section structure

Docs and skills

• refreshed setup, API, CLI, and docs-site content for the current release
• added CLI and Agent Skills documentation
• added the HeadlessX CLI skill install flow for supported AI agents

Included Surfaces

• Website
• Google AI Search
• Tavily
• Exa
• YouTube
• Jobs
• Remote MCP
• Published CLI

Notes

• full Docker now includes yt-engine
• headlessx uses the same backend API and browser state as the dashboard
• Website Crawl remains the only website operator flow that requires Redis and the queue worker