Personal AI agents · one-time setup
Most people never actually run their own AI agent. Setup is the blocker.
I remove it — OpenClaw or Hermes on your cloud in 48 hours. Your keys, your data, flat $200 one-time.
Senior DevOps engineer · 40+ self-hosted LLM and personal-agent deployments
OpenClaw - an open-source personal agent you self-host; it connects to your tools and messaging so an assistant runs on your hardware, not a vendor's.
Hermes - a lightweight, self-hosted agent stack focused on privacy and local or bring-your-own-model LLMs.
40+
Setups delivered
$200
Flat fee
48 hrs
Turnaround
100%
Your infra
The offer
One flat bill - yours. Your cloud, your keys.
OpenClaw — open-source personal agent you self-host; connects to your tools and chat channels so the assistant runs on your server, not a vendor's.
Hermes — lightweight self-hosted agent stack for privacy-first setups with local or bring-your-own LLM APIs.
01
Pick your agent
OpenClaw if you want broad tool and channel integrations; Hermes if you want a lean, privacy-first stack. Both are open and self-hostable - I'll help you choose.
02
Bring your cloud
AWS, GCP, Azure, Hetzner, a Pi in your closet - whatever you're comfortable with. The agent lives on your account.
03
Bring your keys
OpenAI, Anthropic, Gemini, local models - you own the API keys, the spend and the data.
04
I do the wiring
Provisioning, deployment, secrets, auth, persistence, backup setup and a clean handoff doc so you can run it yourself.
Why self-host
SaaS agents rent you access. This runs on your metal.
Typical personal-agent SaaS compared to what Portway sets up for you. Exact SaaS pricing varies by product; patterns are what most tiers follow.
| Topic | Typical agent SaaS | Self-hosted (your cloud) |
|---|---|---|
| Monthly cost | Recurring subscription (often $20–$200+/mo by tier). Billed every month whether you use it or not—seat and message limits per plan. | Cloud/VPS roughly $5–20/mo while running, plus API usage you control. Pay for infra, not a standing SaaS seat tax. |
| Idle months | Still charged. Downgrades may lose features or history. | Stop the VPS or scale to zero when idle; snapshot backups and start again before you need the agent. |
| Usage limits | Message caps, seat limits, and feature gates per plan. | Limits are your API budget and hardware—not an arbitrary SaaS tier. |
| Setup & ongoing fees | Setup folded into subscription; ongoing support tied to your plan tier. | One-time $200 integration with Portway, then you operate it. No mandatory managed hosting or monthly retainer after handoff. |
| Vendor lock-in | Workflows, integrations, and history often tied to one product. | OpenClaw/Hermes on your accounts—your keys, backups, and runbook; move hosts or swap stacks without re-buying access. |
| Data & retention | Unclear what is logged, retained, or used for training; policies change. | Conversations and tool logs on your infra. You choose retention, encryption, and access. |
| Maximum privacy | Usually routes through the vendor's cloud. | OpenRouter zero–data-retention routes, BYOK to OpenAI/Anthropic, or fully local Ollama—your call. |
SaaS can still make sense if you want zero ops. Self-hosting fits when privacy, cost control, or owning the stack matters more than a managed login screen.
From recent setups
Real handoffs, not slide decks.
Anonymized feedback from people who wanted the agent running on their infra — not a managed SaaS trial. 3 setups completed in June 2026 (40+ total delivered).
“Set up my Hermes agent on Hetzner in under 48 hrs. The runbook was solid - I actually know how to restart it without panicking.”
- SWE, Berlin
“I'd been putting off OpenClaw for months. One call, flat fee, agent live on my AWS account. Pass-through costs were exactly what they quoted.”
- Founder, remote
How it works
From kickoff to your agent running in your cloud.
- 01
Onboarding call
We map out what the agent needs to do, walk through cloud and API cost estimates and pick the stack. Costs for infra and APIs are on you - I'll keep them honest.
- 02
Setup & integration
I provision your environment, deploy OpenClaw or Hermes, wire up your API keys, persistence, backups and any tools you need. Flat $200.
- 03
Handoff
You get a working agent, a short runbook and a call to walk through operating it. After that it's yours - no lock-in.
Pricing
One flat fee. No retainer. No lock-in.
Integration
Everything required to get your personal agent live on your infrastructure.
- OpenClaw or Hermes deployed on your cloud
- API keys wired up (OpenAI, Anthropic, local - your call)
- Auth, secrets management, persistent storage and backup setup
- Onboarding call to scope and estimate costs
- Handoff runbook so you can operate it yourself
- Two weeks of follow-up Q&A over email
By scheduling a call with us, you accept our Terms & Conditions.
Borne by you
Pass-through costs
Infrastructure and API costs are paid directly by you to your providers. Typical ranges for personal use:
- Infra: about $5–20/month (e.g. Hetzner, a small VPS, or a modest cloud VM)
- APIs: often $10–50/month for moderate chat and tool use - less if you run local models
We'll map your stack on the onboarding call so there are no surprises.
- Your cloud bill (you own the account)
- API spend (your keys, your budget)
Not included: managed hosting, retainers, or on-call support after handoff (you get the runbook plus two weeks of email Q&A).
Payment
After the onboarding call, you'll receive a Stripe payment link for the flat $200 fee.
FAQ
Questions people ask before booking.
- What if something breaks after the two-week follow-up?
- You keep the runbook and full access to your infra - that's the point. The handoff is built so you can operate it yourself. If you want ongoing help later, we can talk about ad-hoc support, but you're not locked into a retainer.
- Do I need to know how to code?
- No. You should be comfortable creating cloud accounts and pasting API keys. I handle provisioning, deployment, and wiring. The runbook covers day-to-day operations in plain language.
- What cloud do most people use?
- Hetzner and small VPS providers are common for cost and simplicity. AWS, GCP, and Azure work too if you already live there. We'll pick what matches your budget and comfort level on the call.
- Can I switch agents later?
- Yes. OpenClaw and Hermes both run on infrastructure you control. Moving stacks is work, but you're not tied to a vendor - it's your server and your keys.
- What's included in the two weeks of email Q&A?
- Questions about operating what we set up: restarts, backups, rotating keys, light config tweaks. It's not unlimited custom development, but it covers the 'something looks off' period after handoff.
- When do I pay the $200?
- After the onboarding call, once we've agreed on scope. You'll get a Stripe link for the flat integration fee. Infra and API bills stay on your accounts throughout.
Book a call
Pick a time. We'll scope your agent and your costs.
A 30-minute onboarding call to figure out what you want the agent to do, which cloud and API providers fit and roughly what you'll spend. Not ready for a call? Email me first - no pressure.
By scheduling a call with us, you accept our Terms & Conditions.