Running ten stores should not mean ten times the clicking.
datavessel is the operations center for agencies and operators who run many e-commerce stores. One inbox. Rules you set for what fires and how — a fixed step, or an agent that asks first or runs on autopilot where you allow it. A ledger that keeps every receipt, per client.
How multi-store operations should run.
Deterministic where it can be
Most of what an operations team does follows rules you could write on one page. Those rules should run as rules — when · where · then — with no model in the loop and no surprises. A refund under €20 to a repeat customer doesn't need intelligence; it needs to happen.
Agents for the rest, on your rules
The rest is real work: is this chargeback worth fighting, is this ticket really about the order, is this product page losing sales. That's where an agent earns its place — reading the context, doing the work, showing its reasoning — and it asks before money moves unless you've explicitly let it run.
The receipt is the product
Agencies live on trust. Every action has to be answerable: what ran, on which store, for how much, approved by whom. The ledger isn't a log we keep for debugging — it's the thing you hand the client at the end of the month.
Per client, always
A policy belongs to one store. An agent is deployed on one store. Nothing crosses from one client to another — not data, not rules, not mistakes. Scale comes from repeating the setup, not from blurring it.
Your model, your bill
You bring your own Anthropic, OpenAI or Google key. We never mark up tokens or resell inference; the AI bill is between you and your provider. We'll help you set the key up — we won't sit in the middle of it.
Approve from anywhere, chat nowhere
The mobile app does exactly one thing: show you what needs a human across all your stores, with the context and the proposed action, and let you approve or deny. It is not another place to talk to a bot.
Built with the people who run the stores.
datavessel started as agents that did store busywork for a single merchant — the order desk, ad waste, product pages. The first customers who got real value from it weren't single merchants. They were the agencies and operators running several stores for other people, and they all asked for the same things: one place to see what needs them, rules they could trust to run without them, and a clean record per client.
So we rebuilt the product around that job. The inbox, the policies, the ledger and the mobile sign-off app are what an operations team needs when every client runs on a different platform and every mistake has someone else's name on it.
It works on Shopify, WooCommerce and Shopware because that's what a real portfolio looks like. It's built on typed tools and a ledger because that's what makes it answerable.
Founder of datavessel. Ten years across software engineering, data visualization, AI systems and product.
I started datavessel because I think the valuable part of running a business is the judgment — which fights to pick, which customers to keep, what to say no to — and almost none of the day is spent on it. The day is spent on the admin. Agents can carry that weight now, as long as someone stays accountable for every action they take. That accountability, made practical, is the product.
If you run stores for other people and want to talk about how you do it today — write to me directly. I read every email.
Try it on one client's store. Then tell us what's missing.
The product is small and opinionated by design. We add things based on what agencies running real portfolios ask for.