To connect an AI agent to Elevated Product Search, go to Elevated › Product Search › AI Agent, click Install & Activate Agent Connector, then click Open Agent Connector and follow the connection wizard: pick your agent, generate an application password, and paste the setup command into your agent. Once connected, your agent can search products the way your storefront does, change search settings, and manage the search index. Every operation is validated and permission-checked.
Navigate to Elevated › Product Search in your WordPress admin.
Click AI Agent in the plugin’s top navigation, next to Settings.
Agent Connector for WP is the free plugin that exposes Elevated Product Search’s abilities to your agent over MCP. The AI Agent page installs it for you.
Click Install & Activate Agent Connector.
Elevated Product Search downloads a pinned Agent Connector release and verifies its checksum before installing anything. When it finishes, the page shows Agent Connector is active and ready to create a secure connection.
Leave Allow AI agent access switched on. This is the master switch for agent access to Product Search. Turn it off at any time to cut agents off without disconnecting anything else.
Click Open Agent Connector. The Agent Connector screen opens.
Click Get started.
Select the agent you use. Agent Connector shows exact setup instructions for each one; pick Other for universal instructions that work with any MCP-capable agent.
Click Continue.
Your agent authenticates with a WordPress application password, a credential scoped to your account that you can revoke at any time. The agent connects through a local proxy, so the computer running your agent needs Node.js.
Click Generate App Password. If you already created one, click I already have one instead.
The password is shown once. Copy it, or save it to your password manager.
Follow the setup instructions shown for your agent: copy the command, open your terminal, and paste it. Your agent confirms the MCP server was added.
On Mac, the terminal app is called Terminal. On Windows, search for how to open your terminal. The steps are the same after that.
Open your agent and ask:
Are you connected to my WordPress site over MCP? List the Elevated Product Search tools you can use.
The agent should confirm the connection and list the search tools: product search and filter lookups, settings, index management, and diagnostics.
Then try a real request:
Set fuzzy matching in Product Search to normal and rebuild the index.
The agent updates the setting and starts the rebuild; rebuilding the index requires it to confirm the action explicitly.
Elevated Product Search registers its operations as WordPress abilities: the same validated, permission-checked operations the admin interface uses. Agent Connector serves those abilities to your agent over MCP, an open protocol supported by Claude, Codex, Cursor, and most other agents.
The guardrails, in plain terms:
The AI Agent page says WordPress abilities are unavailable. Your WordPress version doesn’t include the Abilities API yet. Update WordPress, then reload the page.
The AI Agent page says Agent Connector is inactive. The connector is installed but switched off. Click Activate Agent Connector.
The setup command fails in the terminal. The connection runs through a local proxy that requires Node.js. Install Node.js from nodejs.org, then run the command again.
The agent connects but has no Product Search tools. Check that Allow AI agent access is switched on on the AI Agent page, then ask the agent to reconnect or refresh its tool list.
You want to cut off access. Switch off Allow AI agent access to block Product Search abilities, or revoke the application password under Users › Profile › Application Passwords to sever the connection entirely.
This is the machine-readable version of this article. Feed it to your agent, or point the agent at the raw Markdown.
Elevated Product Search registers 12 WordPress Abilities under the elevated-product-search/* namespace. Agent Connector for WP serves them over MCP (Streamable HTTP) at:
{site_url}/wp-json/mcp/mcp-adapter-default-serverAuthentication: WordPress application password (Basic auth), typically via the @automattic/mcp-wordpress-remote stdio proxy with env vars WP_API_URL, WP_API_USERNAME, WP_API_PASSWORD.
tools/list on the MCP endpoint returns the exposed tools.mcp-adapter-discover-abilities lists every registered ability; mcp-adapter-get-ability-info returns an ability's full input/output JSON Schema; mcp-adapter-execute-ability executes one.search-products (the same index and filters the storefront widget uses), find-filter-terms, list-searchable-custom-fieldsget-settings, update-settings (partial patch), reset-settings (requires confirm: true)get-index-status, rebuild-index (requires confirm: true), get-rebuild-progress, get-index-scheduleget-environment-status, get-debug-logsThe exact schema of every ability and captured request/response examples: https://elevatedextensions.com/documentation/product-search/managing-with-ai-agent/?format=agent
rebuild-index and reset-settings require confirm: true. The flag is declared in the input schema with true as its only accepted value, so an unconfirmed call fails validation before any handler runs: Ability "elevated-product-search/rebuild-index" has invalid input. Reason: confirm is a required property of input.update-settings takes a partial patch under settings and responds with the full effective settings plus index_rebuilding, updated_keys, and deleted_keys. Verify writes from that response.searchIn* keys) schedules an index rebuild automatically; the response reports index_rebuilding: true.search-products searches the same index the storefront widget uses, honouring the configured include and exclude filters. Its total never exceeds the requested limit; it is not a catalogue-wide count.Allow AI agent access off (option elevated_product_search_agent_access), the abilities are not registered and calls fail with a 404, code rest_ability_not_found, message Ability not found. Ask the human to switch it on; agents cannot.// search-products (input)
{ "query": "dress", "limit": 3 }Response shape:
{
"success": true,
"data": {
"query": "dress",
"total": 3,
"products": [
{ "id": 91, "title": "Black Western Dress", "sku": "", "url": "{site_url}/product/black-western-dress/", "price": "$120.00", "in_stock": true, "type": "simple" }
],
"categories": []
}
}To rebuild the index:
// rebuild-index (input)
{ "confirm": true }Then poll get-rebuild-progress until status is "complete" ({ status, is_rebuilding, products_done, total_products, message, rebuild_id }), and read get-index-status to verify indexed_documents.
