Your idle PCs are the proxy pool.
Vahak turns personal computers, idle laptops, and homelab machines into quiet worker nodes —Vahans — that harvest public pages over your own clean residential ISP. No cloud brute-force, no proxy bill.
Up and running in three commands
One orchestrator, as many Vahans as you own.
How it works
A central orchestrator queues jobs and syncs metadata. Your Vahans do the quiet fetching.
The Orchestrator
A FastAPI server plus web dashboard. Registers workers, queues crawl jobs, and shows live telemetry — status, CPU, memory, active jobs.
The Vahan
A lightweight worker on your own machine. Low-concurrency and polite — it rides a real residential connection, so targets see a household, not a datacenter.
The Edge Buffer
Scraped metadata lands in a local buffer first, then syncs back. A network drop never loses a page you already fetched.
Built for people who read the fine print
Zero-broker orchestration
Worker-to-orchestrator coordination over plain HTTP. No message brokers, no middleware to babysit.
Edge buffering
Every Vahan keeps a local edge buffer that survives network drops and syncs metadata on reconnect.
Fetch-on-demand storage
Pages stay at the edge and are pulled only when a job needs them. The central store holds metadata, not everything.
Bandwidth politeness
Per-node rate limiting keeps each Vahan quiet — low concurrency, respectful intervals, no slamming targets.
| Vahan | Status | CPU | Mem | Jobs |
|---|---|---|---|---|
| vah-01 · homelab | online | 12% | 2.1G | 3 |
| vah-02 · laptop | online | 31% | 3.8G | 2 |
| vah-03 · desktop | online | 4% | 1.4G | 0 |
| vah-04 · mini-pc | syncing | 18% | 2.6G | 2 |
Bring your own idle hardware.
We're onboarding a small group of developers, small teams, and homelab builders into private alpha. Tell us what you're running and we'll be in touch.
No spam. Vahak is built by Loom AI — alpha slots are limited and prioritized by fleet size.