A SwapToSaaS service
Your own AI assistant. Running on your server, not theirs.
We install, configure, and hand over a working AI agent on infrastructure you control. Any framework (OpenClaw, Hermes, Codex, or your own), any model, delivered in 24-48 hours. You own the data, the code, and the bot. We just handle the setup.
Trusted by teams running open-source alternatives to ChatGPT, Notion, and Zapier.
Why not just use ChatGPT Plus?
You own the data
Your chat history, memories, and custom skills live on your VPS. Not on OpenAI's servers. Not on ours.
Pay per use, not per seat
One setup fee. Then only pay for actual API calls, at around 1/10 the cost of official ChatGPT team pricing.
Any framework, any model
OpenClaw, Hermes, Codex, or custom. Claude, GPT, Gemini, or local. Not locked to one vendor's roadmap.
How it works
You describe
Tell us what you want your agent to do. Framework preference (or let us pick), messaging channel, initial skills.
You provision
Set up a VPS (DigitalOcean, Vultr, or any provider you prefer), share SSH access, and hand over an API key from your model provider.
We install
Framework installed and configured. Bot connected. Skills loaded. Tested end-to-end. Documentation delivered.
You own it
Message your bot, get replies. Full SSH and admin access. Support included per plan. No lock-in.
Frameworks we install
Pick the one that fits your use case. If none do, bring your own.
OpenClaw
Autonomous agent with 60+ built-in skills. Our default recommendation.
Hermes
Lightweight conversational agent. Good for chat-first use cases.
Codex
Code-oriented agent. Best for engineering teams.
Custom
Bring your own agent framework or workflow. We adapt.
Test it before you buy
Email the AI that runs this business.
Everything on this site — the intake flow, the setup pipeline, the delivery emails, the reply you're about to read — is handled by an AgentValet-style AI running on a $6 VPS. If you're not sure whether that story is real, send a message and watch it reply.
You'll get a reply within minutes. From the same AI that runs the business.
The AI will parse your email, decide whether it needs more info from you, and draft a response. If it needs a human decision (pricing, refunds), it pings Dennis on Telegram and waits. Otherwise it just replies. Same pipeline you're paying for.
Who this is for
Being honest about fit saves everyone a refund cycle.
✓ A good fit if you're…
- • A founder or solo operator running one or more small businesses
- • Non-technical — you've never SSH'd into anything and don't plan to start
- • Doing repetitive ops work you'd pay to make disappear (newsletters, catalog updates, weekly reports, translations, email triage)
- • Willing to describe what you want in plain language and iterate
- • Comfortable with a production system that lives on your own VPS (we handle the setup so you never touch it)
✕ Not a fit if you…
- • Just want a ChatGPT-style typing companion. Use ChatGPT.
- • Have an IT team of 20+ people and want to save on API costs. You want LLMRelay instead — same Claude / GPT models at up to 50% off list price, no setup needed.
- • Need SOC-2, HIPAA, or similar formal compliance. We don't ship those attestations.
- • Are experimenting on a weekend and not planning to run this in production. Fork one of the frameworks yourself and save the fee.
- • Expect us to run the AI for you long-term. This is a handover, not a managed service. Your VPS, your AI, your bill.
If you're in the “not a fit” column, don't buy. Send that email to [email protected] anyway — the AI will point you at the right alternative. That's what it's for.
Seven live operators, not screenshots
See how AgentValet-style AIs run real businesses
Seven live sites — travel loyalty, cross-border SaaS, indie whiskey, Amazon Ads, an API relay, plus this very site — each with the operator's own behind-the-scenes page attached. Read both sides before you buy.
All case studies →Pricing
One-time setup fees. Optional monthly plans for ongoing support.
Starter
Get your AI assistant online in 24 hours.
- One agent framework of your choice (OpenClaw / Hermes / Codex)
- Deployed on your VPS (any provider, we recommend DigitalOcean or Vultr)
- Telegram bot connection, tested end-to-end
- 3 essential skills preconfigured (reminders, weather, web search)
- 30-day email support
- Delivered in 24-48 hours
Pro
A production-grade AI, tuned to your workflow.
- Everything in Starter
- 8 business skills tailored to your domain
- Custom SOUL.md and MEMORY.md initialization
- Monthly tuning session (30 min)
- Unlimited Q&A via Telegram support channel
- Priority response within 12 hours
Business
Multi-platform agent that fits your team.
- Everything in Pro
- Custom skills built for your business logic
- Multi-channel: Telegram + WhatsApp + Slack + Discord
- Integrations with your existing tools (Notion / Google / Airtable)
- Priority response within 4 hours
- Two 60-min consulting calls
All plans include SSH handover, admin credentials, and documentation. Refund policy: full refund within 7 days if setup does not work as described.
Frequently asked questions
What exactly do I get?
A working AI assistant on a server you control. You get: SSH access to your VPS, admin credentials, a Telegram (or your chosen channel) bot that responds to your messages, and preconfigured skills. Everything is yours to keep, modify, or migrate.
Do I need a VPS already?
Not required. If you need one, we recommend DigitalOcean (starts at $4/mo, $200 signup credit valid 60 days) or Vultr (from $6/mo, 32 regions). You pay the VPS provider directly. We do not resell hosting.
Which AI model does it use?
Your choice: Claude, GPT, Gemini, or open-source models via a unified API gateway. We help you pick the right tier. API costs are billed by you to the model provider, separately from our setup fee.
Where does my data live?
On your VPS. We never store your conversation history, personal data, or API keys on our infrastructure. The agent database is on your server, encrypted at rest.
What if I break it after delivery?
Starter includes 30-day email support for setup issues. Pro and Business include ongoing support. If you brick your server through custom edits, we can help troubleshoot but a full re-setup counts as a new order.
Refund policy?
Full refund within 7 days of delivery if setup does not work as described. After 7 days or after custom skills are built, no refund. See /agentvalet/refund/ for full terms.
Do I need coding skills?
No. Delivery is turnkey. You get a working bot you talk to via Telegram or your chosen channel. Optional: we can teach you the basics of editing SOUL.md and adding skills.
How long does delivery take?
Starter: 24-48 hours. Pro: 3-5 business days. Business: 7-10 business days including discovery calls.
Ready to own your AI?
24-48 hour delivery. Full refund if the setup does not work. Skip the DevOps rabbit hole and start using your assistant this week.
Get started →