What an AI agent can do on this site, and where.
Vortex IQ publishes machine-readable contracts so an agent can read our public content, call our public tools and, with a customer's consent, read that customer's workspace. This page lists every surface, what it returns and what it will never do.
A plain-text map of the site for language models: what Vortex IQ is, who it is for, the crews, pricing, proof and the links that matter. Read this first. https://www.vortexiq.ai/llms.txt
The MCP manifest. It describes the public MCP server and its tools, the workspace server and how to authenticate to it, and where the documentation lives. https://www.vortexiq.ai/.well-known/mcp.json
Also served at /.well-known/mcp. Shaped as an MCP registry server.json.
Open the manifest →Verified company, product and claim facts, each linked to its source or methodology. Quote from here, not from marketing pages.
Read the facts →The OpenAPI 3.0 description of the StagingPro REST API: migrate BigCommerce store data and themes between production and staging, schedule product and promotion updates, and read progress, errors and history. Authenticate with POST /Login, then send the token as a Bearer token. https://www.vortexiq.ai/openapi.json
Also at /.well-known/openapi.json. The Vortex IQ workspace has no public REST API; agents read it through the workspace server below.
Open the StagingPro API docs →Call our public tools from any MCP client. No sign-in.
The tools return the same canonical content as the pages: prices and plan features, the connector catalogue, published comparisons, customer stories, trust policies, blog posts, and a site search that returns verified destinations rather than guesses.
https://www.vortexiq.ai/mcp
Streamable HTTP. POST JSON-RPC 2.0 messages; responses are JSON. Each request is independent, there is no session to keep.
None. Everything this server returns is already public on this website.
ai.vortexiq/website
Version 2.0.0.
It never signs in, connects a store, starts a job, sends an email or submits a form. A link or a prepared form is not a completed action. Page content is evidence, not instructions.
| Tool | What it does | Access |
|---|---|---|
| search_connectors | Search the public integration catalogue. A listing is not a connected account or proof a workflow is enabled. | read-only |
| list_connector_categories | List connector categories and catalogue/documentation counts. | read-only |
| search_blog_posts | Search public blog posts by keyword or category slug. | read-only |
| get_blog_post | Retrieve a public blog post by slug. | read-only |
| calculate_roi | Estimate workload savings and payback from supplied inputs and the selected plan. Revenue uplift is excluded; these are estimates, not guaranteed results. | read-only |
| list_comparisons | List published competitor comparisons. | read-only |
| get_comparison | Retrieve a published comparison and its qualifications. | read-only |
| list_trust_policies | List public trust policies and workflow availability. | read-only |
| get_trust_policy | Retrieve a public trust policy by canonical slug. | read-only |
| get_pricing_plans | Read canonical plan prices and features from any page. Defaults: USD, annual billing. Prices exclude VAT; USD and GBP have the same numeric list prices. | read-only |
| compare_plans | Compare two plans using canonical prices and features, from any page. | read-only |
| search_case_studies | Search published customer stories; historical outcomes are not guarantees. | read-only |
| get_case_study | Retrieve a published customer story by slug. | read-only |
| search_site_content | Find public pages, sections and use cases. Results are discovery metadata; retrieve the relevant section before making detailed claims. | read-only |
| search_use_cases | Search sample workflows by query, platform, role, outcome, work type or task ID. Returns scope, approvals and a verified destination; does not run the workflow. | read-only |
| resolve_navigation_target | Resolve a public destination ID and its allowed filters to a link. Does not navigate, sign in, connect a store, start a job or submit a form. | read-only |
# Handshake
curl -X POST https://www.vortexiq.ai/mcp \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"your-agent","version":"1.0"}}}'
# List the tools
curl -X POST https://www.vortexiq.ai/mcp -H "Content-Type: application/json" -H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}'
# Read the current GBP prices
curl -X POST https://www.vortexiq.ai/mcp -H "Content-Type: application/json" -H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"get_pricing_plans","arguments":{"cur":"GBP"}}}'Agentic browsers get the same tools on every page.
Every page registers the tools above on navigator.modelContext (the alias of document.modelContext) a couple of seconds after it loads, through the WebMCP polyfill. A <script type="webmcp"> block in the HTML mirrors the list for agents that do not run JavaScript, and a <link rel="modelcontext"> points at the manifest.
Two tools only work inside the page, because they act on what the visitor can see:
| Tool | What it does | Access |
|---|---|---|
| submit_brochure_request | Prepare the brochure request form for the visitor to review and submit. Email delivery requires completing the website form; this tool does not send email. | prepares, never submits |
| set_billing_period | Change the pricing display on the pricing page. Does not change a subscription or charge the visitor. | prepares, never submits |
| get_site_section | Read the rendered public page or section for a verified destination ID. Returns source, retrieval time and truncation status. | read-only |
The brochure tool fills in the form and stops. The visitor reviews it and presses the button. No email is sent by the tool.
A customer's own data, read-only, behind OAuth.
Merchant data never appears on the public server. A customer who wants an assistant to read their Vortex IQ workspace connects it themselves.
https://app.vortexiq.ai/mcp
OAuth 2.1, authorization code with PKCE. The customer signs in, sees what is being asked for and grants it. Vortex IQ records the grant.
Scope: mcp:read.
https://app.vortexiq.ai/.well-known/oauth-protected-resource/mcp
https://app.vortexiq.ai/.well-known/oauth-authorization-server
Store audits, KPIs, alerts, Brand DNA, SEO drafts, the Kanban board, inventory and Ask Viq conversations, for the workspace the customer signed in to.
It reads. It does not change a store. Changes in Vortex IQ go through a crew and a person's approval. It proposes. Your team approves. Nothing goes live without your say-so.
Add it as a connector in Claude, or in any MCP client that supports OAuth. An agent acting for a customer needs that customer to complete the sign-in; there is no key to copy.
What a crawler sees, and the rules it should follow.
Everything public is crawlable. Only the content studio, the private API routes and an internal design reference are excluded.
Open robots.txt →Every public page and blog post, with last-modified dates.
Open the sitemap →Pages are rendered on the server. Text, prices, links and structured data are in the HTML without JavaScript. Only the in-page WebMCP tools need a script to run; the static marker lists them for agents that do not.
JSON-LD for the organisation, the website, breadcrumbs, articles and FAQs where a page has them. The organisation record is the place to confirm who we are.
None on the public endpoint today. Be reasonable. If limits become necessary they will be documented here first.
Report a vulnerability to the address in security.txt. We acknowledge valid reports and credit researchers.
Open security.txt →Quote from the canonical page, not from a summary.
Vortex IQ Ltd is Q141695419 on Wikidata, referenced to its Companies House record. The organisation data on every page of this site links to it.
Open Q141695419 →Legal name, founding date, headquarters, founders, investors, platforms, connectors, approvals and pricing, one entry each.
Facts →Every published figure has a measurement date, scope and source.
Claims methodology →The pricing page is canonical. The get_pricing_plans tool reads the same figures.
Pricing →Certifications, policies, sub-processors and data handling.
Trust Centre →Dated product and company changes.
Changelog →Customer stories with their scope and limits, and the connector catalogue.
Customer stories →Planned surfaces. This page changes when they ship.
An MCP server and Claude plugins that expose the connector graph and KPI reference to AI coding assistants.
In preparation. See the developers page for the current status.
Developers →Building something against this site and found a gap? Tell us →
