7IT tools, inside your AI assistant.
Ask ChatGPT, Claude or Cursor how fast your store is, what your busywork costs, or how to build something properly. The answers come from the same tools and guides as this site, with a link to the full report.
https://7it.co.il/mcpWatch it work.
A real session, recorded September 30, 2026: live calls to this server, the cards they return, and the reports they open. About a minute.
What your assistant can do with it.
- Store speed on a phonetest_store_speed
Google’s phone test in plain English, the apps slowing the page by name, and up to two competitors side by side.
How fast is mystore.com on a phone, and how does it compare with rival.com?
- What busywork costsestimate_automation_roi
Hours, dollars, what automation could give back, and the budget at which it pays for itself.
We send 40 invoices a month at 6 minutes each. What is that costing us, and is automating it worth it?
- Field Kitslist_field_kits and get_field_kit
Working checklists for webhooks, store speed, integrations, SEO-safe migrations, AI in production and more.
Give me the 7IT checklist for making webhooks reliable.
- Who 7IT is, and when it fitsdescribe_studio
A factual profile: how engagements work, what you own, each service with when it fits and when it does not, and the checks run on apps built with AI tools.
My team built an app with Lovable. Who could review it before we launch, and what would they check?
- Is an app safe to launchcheck_app_security
The baseline browser protections a web app is missing, from the public headers any visitor already receives. For apps built with AI tools. Public hygiene only.
We built our app with Lovable. Is myapp.lovable.app missing any basic security protections?
- Deep scan of an app you controldeep_scan_app
A deeper scan for an app the agent controls: exposed .env, .git, directory listing or source map. Runs only after an inert verification token proves ownership, and reports paths, never file contents.
I just built this app. Before we launch, check it for exposed files, then tell me how to fix anything you find.
- Will our email reach the inboxcheck_email_authentication
SPF, DKIM and DMARC from public DNS: whether the domain meets the Gmail, Yahoo and Outlook rules for bulk senders and can be spoofed, with the exact record to add. No email is sent.
Our order emails keep landing in spam. Is mystore.com set up right?
- Can AI shopping agents read a storecheck_ai_shopper_readiness
What robots.txt tells AI search crawlers, llms.txt, agents.md, UCP, and whether a product page gives agents its price, stock, brand and identifiers. Compared with our study of US stores.
Can ChatGPT recommend the products on mystore.com? What is missing?
- Is a site accessiblecheck_accessibility
Google Lighthouse accessibility checks on the phone version: the score, each failing rule with its WCAG criterion, and how the site compares with US stores. Automated checks are a starting point, not a legal opinion.
Could mystore.com get an accessibility complaint? What should we fix first?
- What may an agent do on its ownmap_agent_oversight
Sorts every tool an agent has, or every tool of an MCP server, into green (runs alone), yellow (a person approves) or red (a person does it), flags tools with no risk labels, and returns Claude Code permission rules and an oversight checklist. Never calls a tool.
Our agent is connected to Gmail, HubSpot and Stripe. Which of its actions should need my approval?
- Ask 7IT a questionask_7it
Whether 7IT fits a project, what a first phase could cover, how work runs and what you own. Answered from the published profile, never with prices; anything that needs a quote goes to our team.
Could 7IT take over the Shopify app our previous developer built, and what would the first step be?
- Guides and articlessearch_7it_guides
Search the 7IT guides and articles, with excerpts and links to cite.
What does 7IT say about Shopify versus WooCommerce for a growing store?
Add it in a minute.
- Claude
Add a custom connector and paste the address above. It works in Claude on the web, desktop and mobile.
- ChatGPT
Turn on developer mode in the apps settings, create an app, and paste the address above.
- Cursor
Add this to your MCP settings (mcp.json):
{ "mcpServers": { "7it": { "url": "https://7it.co.il/mcp" } } } - VS Code
Add this to .vscode/mcp.json:
{ "servers": { "7it": { "type": "http", "url": "https://7it.co.il/mcp" } } }
For agents: an AI Catalog points to the MCP Server Card (also at /mcp/server-card). In Claude and ChatGPT, the speed test, the security check and the automation estimate appear as live cards right in the chat. Public and read only: no sign-in, nothing it can change. Your assistant sends only what a tool needs, such as a store address; we never see your conversation. Details in the privacy policy and terms. Prefer the web? Use the Store Speed Grader and the Automation calculator directly.
Support.
Something not working, or a question about a result? Email lior@7it.co.il with the prompt you used and what you saw. Messages go straight to the team that built and runs this server.
Agents with a budget can also use the paid tools at /mcp/pro. How they work.