Steward handles your daily work today. Scout joins the lineup for research and deep analysis.
Your daily operator
Research & deep work
Reads, sorts, drafts replies. You just approve.
Books, blocks, reschedules. Zero back-and-forth.
Visits sites, pulls data, drops notes into your CRM.
Writes, schedules, publishes across every channel.
Follows up on late payments so you never have to.
Sorts downloads, renames folders, finds anything.
DIY hosting is cheap until you count the Sunday you spend on Docker, TLS, and webhooks — or wake up to a public URL with no login.
Your own box. BYOK keeps model spend on your Anthropic or OpenAI account. A $5–20/mo VPS can look cheaper on paper — if you have the time.
Researchers have found tens of thousands of exposed self-hosted AI assistants globally, many without authentication — not because people are careless, but because secure defaults are hard. We ship auth on, updates, and private networking so you do not have to become a part-time SRE.
Every plan here is BYOK: we charge for your robot running 24/7; your LLM provider bills usage separately. Light use is often ~$15–40/mo. Heavy agentic loops can spike fast — we help you set caps so one runaway task does not become a surprise bill.
Hosting is the easy part. Keeping it safe and online is the work. We are $29–49 for a personal assistant that runs 24/7 — same BYOK model, no weekend hardening.
Sign up and choose Starter or Pro. Steward is included — inbox, calendar, and documents ready to go.
Paste your LLM key, name your assistant, and connect optional channels in Settings.
Talk to your robot in dashboard chat. Connect Google Calendar, Tasks, Gmail, and Drive from Modules when you are ready.
Comparable to self-hosting a VPS ($5–20) plus your Sunday — without the security homework. Your AI provider bills tokens separately (BYOK).
Your personal assistant drafts inbox replies, chases invoices, and preps your calendar — you approve before anything sends. Starter ($29) is lighter compute; Fleet ($149) is waitlist-only until multi-robot ships.
Your personal assistant for inbox, calendar, invoices, and documents — running 24/7.
Talk to your robot in the dashboard — web today; iOS and Android coming soon.
Drafts and multi-step jobs pause before email, posts, or side effects — you say go.
Auth on, updates, private networking — not a public gateway on a $5 VPS.
Pay, paste your LLM key, name your robot — we provision and health-check in ~90 seconds.
Telegram, SMS, Discord, and more in Settings — connect only if you want them.
API spend stays on your provider; token safety caps protect against runaway loops.
One robot, always on
1 vCPU · 1 GB RAM
More headroom for busy agents
2 vCPU · 2 GB RAM
Team tier — join the waitlist
Target: 3× 2 vCPU · 2 GB
Your robot only runs as fast as its CPU and memory. Cloud APIs answer quickly; skills, memory graphs, and channel bridges still need local compute. Higher tiers buy more headroom — not a different logo.
| Starter $29 | Pro $49 | Fleet $149 | |
|---|---|---|---|
| Instances | 1 | 1 | 3 (waitlist) |
| vCPU (each) | 1 | 2 | 2 |
| RAM (each) | 1 GB | 2 GB | 2 GB |
| Token safety cap | 500k/mo | 2M/mo | 5M/mo |
| Best for | Light daily tasks | Busy assistant + chat | Teams & parallel robots |
Dashboard chat — talk to Steward in the dashboard. Documents, calendar, and email modules on Pro.
iOS & Android apps — same chat, simpler UX. Scout model for research and deep work.
| Usage | Hosting | LLM (BYOK) | Typical total |
|---|---|---|---|
| Light daily chat | $29–49 | ~$15–25 | ~$44–74 |
| Moderate automation | $29–49 | ~$30–60 | ~$59–109 |
| Heavy agentic loops | $49 | ~$80–150+ | Use token caps |
API cost drivers: model choice, long threads, cron digests, tool loops. Token safety caps protect against runaway bills.
Cancel anytime — service through end of billing period. 7-day money-back on first subscription if we cannot get you live.
Hosting: $29 / $49 — less than a billable hour of a VA. Fleet $149 is waitlist-only.