What people build on Machineuse
Anywhere you need many real, isolated browsers running concurrently — scraping, testing, agents, and rendering — Machineuse runs them as a self-hosted fleet.
Machineuse is used for web scraping at scale, browser and E2E test farms, agentic browser automation for AI agents and MCP tools, andscreenshot/render services — any workload that needs many isolated browsers scheduled across worker nodes.
Web scraping & data collection at scale
Run large crawling and extraction workloads across many nodes with per-session isolation.
Browser & end-to-end test farms
Parallelize browser test suites across a fleet without a fragile central grid.
Agentic browser automation (AI agents & MCP)
Give AI agents and MCP tools a scalable pool of isolated, disposable browsers.
Screenshot & render services
Turn URLs and HTML into images or PDFs at volume, on isolated browsers.
Web scraping & data collection at scale
Run large crawling and extraction workloads across many nodes with per-session isolation.
Large scraping jobs need many real browsers running concurrently, each with its own cookies, storage, and fingerprint. Sharing one browser across targets leaks state and gets you blocked; running them all on one host melts it.
Machineuse places each scraping session in its own systemd-nspawn container and schedules it onto a worker with spare capacity. Sessions stay isolated from one another, and snapshot dormancy lets you keep far more logical sessions than you could hold live at once — reviving a session only when you need to crawl again.
Browser & end-to-end test farms
Parallelize browser test suites across a fleet without a fragile central grid.
CI pipelines want to run hundreds of end-to-end tests in parallel, each in a clean browser. Traditional Selenium/Playwright grids funnel this through a central hub that becomes a bottleneck and a single point of failure.
Machineuse spreads test sessions across worker nodes with load-aware scheduling and no central broker. Each test gets a fresh, isolated container, so tests never contaminate each other. Add worker nodes to add test throughput.
Agentic browser automation (AI agents & MCP)
Give AI agents and MCP tools a scalable pool of isolated, disposable browsers.
Computer-use agents need real browsers they can drive safely — ideally disposable, isolated, and available on demand. Handing an autonomous agent a shared browser is risky; provisioning a fresh VM per agent is slow and wasteful.
Machineuse gives each agent an isolated browser in its own container, scheduled onto available capacity. The project includes MCP-oriented tooling for programmatic control, so agent frameworks can request, drive, and release browsers as part of a fleet.
Screenshot & render services
Turn URLs and HTML into images or PDFs at volume, on isolated browsers.
Rendering services — screenshots, social cards, PDF generation — are bursty and CPU-heavy. Keeping browsers warm for every possible tenant is expensive; cold-starting one per request is slow.
Machineuse schedules render jobs onto workers by live load and uses snapshot dormancy to keep warm sessions cheaply, reviving them on demand. Per-session isolation keeps one tenant’s render from touching another’s.
Teams that run Machineuse
Scraping & data teams
Reliable, isolated browser pools that scale horizontally as crawl volume grows.
QA & test engineering
Parallel E2E execution across a fleet with clean, reproducible browser environments.
AI & agent builders
On-demand isolated browsers for computer-use agents, exposed via MCP tooling.
Platform engineers
A self-hostable, brokerless system to operate — no SaaS bill, no central broker to babysit.
See how it runs
Read the architecture, compare it to the alternatives, or check the FAQ.