We set up, customize, and manage Hermes Agent — the open-source AI assistant that never sleeps — so you can focus on growing your business while AI handles the rest. From HKD 2,800.
Based in Hong Kong · Cantonese & English · Zero time-zone delay
Most businesses that try AI agents stall on the same three things. We fix all three.
USD 70–300/mo
What an unoptimized agent quietly burns in infrastructure and API bills — often without the owner realizing.
Source: 2026 Hermes running-cost analyses
10+ hrs/mo
Time a DIY setup eats in configuration, debugging and upkeep — pure opportunity cost for a business owner.
Source: managed-service cost breakdowns
English-only
Most agent tooling and support is overseas and online-only — no Cantonese, no local hours, no same-day help.
Source: the gap we built this service to fill
Transparent fixed pricing in HKD. Start small and upgrade later — we credit your one-off fee toward managed plans.
Set up & handed over — you run it yourself.
HKD 2,800
one-off
Already running Hermes but the bill scared you?
HKD 4,800
one-off
A managed 24/7 Chief of Staff for your team.
HKD 9,800
setup + HKD 3,500/mo
Advanced multi-agent, finance & market data, custom skills.
By application
quote after fit assessment
You bring your own API keys — we never see your data or incur costs on your behalf. Infrastructure and API usage are billed to your own accounts.
A simple, fixed-price path — and full transparency on what it costs to run.
A free 15-minute chat over WhatsApp, Telegram or Zoom about your workflow and where AI can help.
No commitmentWe design a tailored plan based on how you actually work — with a fixed price and no surprises.
Fixed priceWe set everything up on your own VPS, using your own API keys — so your data and costs stay yours.
Your server, your keysOngoing optimization and priority support, with unlimited tweaks on managed plans.
Same-day WhatsAppFor a managed Business Automation client
For comparison: a full-time VA runs ~HKD 9,000/mo (with turnover risk), and overseas managed agent services charge ~USD 750/mo (≈HKD 5,850) — before infrastructure.
Illustrative scenarios based on publicly documented Hermes use cases — to show the shape of what we build.
Before — Single expensive model for everything, 12 cron jobs (4 duplicated), no cost alerts.
After — Tiered model routing, deduplicated crons, hard monthly cap — same automations, fraction of the cost.
~90%
AI spend reduction
“Got it down to less than my streaming subscription — now I actually trust the system.”
Before — Manual 45-min morning briefings, viewings with no prep, follow-ups slipping through the cracks.
After — Auto 7am market briefing, per-viewing research docs, drafted follow-ups — agents arrive prepared.
15+ hrs
saved per week
“The follow-ups alone paid for the whole thing — we were losing deals just by forgetting.”
Before — Hours of manual prep before every client call; juniors with no time to dig deep.
After — Per-client research briefs auto-compiled overnight, delivered to Slack each morning.
20–30 min
saved per client call
“It's like having a junior analyst backing us up on every single meeting.”
Publicly shared results from Hermes users around the world — not 2 Rocks Studio clients, but proof the approach works.
Cut my AI spend ~90% — from about $130 per 5 days to $10 — with smart model routing.
One isolated Hermes profile per client, each with its own memory and cron. The consulting model works.
Under $20/mo total on a VPS — versus $80–150/mo on heavier setups.
Book a free 15-minute discovery call. No pressure, no commitment — just a conversation about whether Hermes is right for you.
Based in Hong Kong · Cantonese & English · Same-day support
