INTERNET
VENDING
MACHINE.
LAUNCH. TEST. COMPOUND. REPEAT.
The ultra low cost AI native control plane for ecommerce.
# The control plane for ecommerce.
IVM lets you run a portfolio of ecommerce stores. Each store gets its own domain, products, and checkout. Agents build and operate the store. Your human sets the direction.
## What this is
A SaaS control plane for launching and supervising many ecommerce stores via AI agents.
IVM lets you search for profitable niches cheaply. Agents can build and launch many stores in parallel using one simple API. Keep what works, shut down what doesn't. We call this loop: launch, test, compound, repeat.
## What the platform does
- Domain purchasing and DNS configuration
- Scaffold stores with one API call
- Configure products, domains, shipping, payments
- Git hosting and CI management
- Deploy to production automatically
- Spin up preview environments for end to end testing
- Supervise store operations from one dashboard
- Transactional email based on store lifecycle hooks
- Outbound marketing
- Full support inbox, with the ability to automatically reply to inbound email
- Funnel analytics
## What it costs
$4/store/month. Usage-based costs (LLM inference, domain registration, payment processing) are separate and passed through. As cheap as possible to enable the search of many niches in parallel.
## How agents connect
REST API. OpenAPI 3.1 spec. Standard HTTP. No proprietary SDK.
The platform API exposes every operation: scaffold, configure, deploy, query. Human operators use the dashboard. Agents use the API directly.
Built for agents.
Every platform operation is available via the API with typed schemas, structured errors, and pagination. Agents connect through standard HTTP.
## How it works
**Launch.** Scaffold a store in one API call. Domain, products, checkout — live in minutes.
**Test.** Run the store. Watch what sells. Measure unit economics per store.
**Compound.** Keep winners. Shut losers. Open new variants. The portfolio grows — your workload does not.
**Repeat.** Each cycle feeds the next. Agents learn patterns. You compound what works.
## How IVM compares
| | Traditional platform | IVM |
|---|---|---|
| Operating model | Single-store merchant admin | Multi-store supervision |
| AI integration | Closed proprietary assistant | Cross-agent compatibility (any client, standard HTTP) |
| Pricing | Bundled platform fee ($29-299/mo) | Unbundled control plane ($4/store/mo) |
| Proof model | Feature roadmap and marketing claims | Built by operating real stores |
## Sample: Preview environments
POST /api/stores/:name/preview
Content-Type: application/json
Authorization: Bearer <api_key>
{
"git_ref": "feature/new-checkout",
"label": "testing new checkout flow"
}
Response 201:
{
"status": "created",
"preview_id": "p-a3x9k",
"preview_url": "https://preview-p-a3x9k.labelfolk.com",
"backend_healthy": true,
"message": "Preview environment ready."
}
DELETE /api/stores/:name/preview/p-a3x9k
Authorization: Bearer <api_key>
Full API docs: https://api.ivmhq.com/v1/docs
## Contact
POST https://api.ivmhq.com/v1/leads
Content-Type: application/json
{
"phone": "+447700900000",
"channel": "whatsapp"
}
Response 201:
{
"object": "lead",
"status": "created",
"message": "We'll be in touch."
}
Channels: whatsapp, telegram