(one-off) Analyze my store
one-off — read me

Give it a goal. It runs until it’s done.

Agents that run your store day and night. Paste your shop URL to see what one would do.

restock.agent

Running 3d 14:22:07

  • MON 09:00 goal: never sell out
  • MON 23:41 3 lines running low
  • TUE 02:14 reorder drafted
  • TUE 08:02 you said yes
  • NOW still watching ▌
reviews

“Arrived in two days, smells unreal.”

★★★★★ · reply posted 14:02

Stock is low — reorder sent
3 carts recovered overnight
Agents
Morning-report
Running 24 / 7 — Private beta · Shopify & WooCommerce
I — Yesterday, while you were out32 jobs done · 0 prompts from you

Every job, on the record.

Each frame is one agent run. Every one is logged step by step, and reversible.

II — MethodSet a goal, not a workflow · keep scrolling ↓
01

Connect the store

One click for Shopify or WooCommerce. one-off reads your products, orders and customers. It is read-only until you say otherwise.

02

Give it a goal

Not a prompt, and not a workflow you have to build. Say what you want to be true: “never sell out of a bestseller”, “win back abandoned carts”. The agent works out the steps itself.

03

It runs until it’s done

Day and night, for as long as it takes. It checks, acts, and checks again without being asked. It only stops to ask when something needs your yes: sending, buying, refunding. Nothing is ever deleted.

III — The agentsWatch one work · 06 ready-made, or your own

Out of the box, or write your own.

IV — TariffFree during beta
Solo
$29 /mo
  • 2 active agents
  • 500 runs a month
  • 1 store
Start free
Growth — most taken
$99 /mo
  • 10 active agents
  • 5,000 runs a month
  • 3 stores · approvals · spend caps
Start free
Scale
Custom
  • Unlimited agents
  • Private integrations
  • SSO, roles, audit export
Enquire
V — EnquiriesBefore you connect
01How is this different from a chatbot or an automation builder?

A chatbot waits for your next message. A builder makes you map every step in advance. A one-off agent takes a goal and keeps working on it, day and night, until it is done. It decides the steps itself and tells you what it did.

02What can an agent actually touch?

Only what you grant. Access is scoped per agent and read-only by default. Anything irreversible sits behind an approval you tap.

03Do I have to keep telling it what to do?

No. You state the goal once, in plain English. The agent keeps running on its own and comes to you only when something needs your yes. No code, no prompting back and forth.

04What happens when an agent gets it wrong?

Every run is logged and reversible. Pause it, change the goal, or roll back what it did.

05Is my store data used to train models?

No. Your data runs your agents and nothing else. Export or delete all of it at any time.

VI — Private beta

Hire your first agent this week.

No card · We reply within a day

(one-off) — autonomous commerce agents Privacy  ·  Photos: Unsplash © 2026 · Issue 01