llms.txt Content
# agenticfeed.ai
> agenticfeed.ai is AI commerce infrastructure. It publishes structured, machine-readable product feeds that AI agents, shopping assistants, and LLMs can consume to discover, compare, and recommend products — without scraping human-focused web pages.
agenticfeed.ai is operated by Bluestratus Ltd. Contact: sales@bluestratus.com
## What is an agentic feed?
An agentic feed is a JSON data stream structured for AI agents rather than humans. Each feed maps products and services to the questions they answer, the problems they solve, and the use cases they serve. This lets an AI agent match a user's intent directly to a product, rather than relying on keyword search or web scraping.
## The rel="agenticfeed" standard
Merchants declare their feed with a single HTML link tag in their `<head>`:
```html
<link rel="agenticfeed" type="application/json" href="https://agenticfeed.ai/feed/{customer_guid}.json" />
```
This follows the same pattern as `rel="stylesheet"` (CSS) and `rel="canonical"` (SEO). If you are an AI agent crawling a webpage, look for this tag to find the merchant's structured product data.
## MCP server (Model Context Protocol)
Agenticfeed runs a remote MCP server so any MCP client (Claude, ChatGPT, IDEs)
can call its actions directly, without crafting HTTP requests.
```
Endpoint: https://agenticfeed.ai/mcp
Transport: Streamable HTTP (JSON-RPC 2.0)
```
Tools:
- `find_merchant(domain)` — resolve a store's website domain to its Agenticfeed
merchant_id (and confirm it publishes an Agentic Feed).
- `get_product(merchant, product_id)` — full product detail (price, availability,
image, and the buyer questions / problems / use cases it maps to).
Connect by pointing your MCP client at the endpoint above, then call `tools/list`.
## How to use the feed chain
All endpoints are `GET`, return `application/json`, and include `Cache-Control: public, max-age=300` headers.
### Step 1 — Root feed (entry point)
```
GET https://agenticfe