Armada, a K8s orchestrator for bots and scrapers

Name: Armada, a K8s orchestrator for bots and scrapers
Availability: InStock
Author: armada_services

by armada_services·Mar 2, 2026·2 points·0 comments

Visit Project View on HN

AI Analysis

●●●BangerSolve My ProblemSlick

K8s scraper framework with built-in physics-based mouse curves and fingerprint spoofing, finally.

Strengths

•Ships with Fantomas (nodriver wrapper), WindMouse, and randomized keystroke libs—solves the actual hard problems, not just orchestration.
•RabbitMQ + Redis + dedicated proxy/fingerprint services mean zero code changes to scale from 1 to 100+ workers.
•AGPL with comprehensive docs, live demo, and working bootstrap scripts—production-ready, not theoretical.

Weaknesses

•Azure SQL dependency in bootstrap scripts limits appeal; no lightweight alternative (SQLite, PostgreSQL) shown.
•Steep setup: K8s cluster, Docker, multiple services required. Overkill for teams under 10 workers.

Post Description

I got tired of gluing together proxy rotation, fingerprint spoofing, job queues, and deploy scripts every time I needed to run browser automation (bots/scrapers) at scale. So I built an orchestration layer for it.

You write a Python script (Playwright, Selenium, or nodriver), add a JSON config, and Armada distributes the work across K8s pods. Jobs go through RabbitMQ, per-pod configs are seeded via Redis, proxies and fingerprints are handled by dedicated services. Scaling from 1 to 100+ workers needs no code changes.

It also ships with Fantomas, a browser automation lib on top of nodriver, physics-based mouse curves (WindMouse), randomized keystrokes, and an XDO mode that fires OS-level events via xdotool.

Code: https://github.com/armadasvc/armada (AGPL v3)

Docs: https://armada.services/docs

Reach out anytime.