NorthCinder

Your shopping agent should work for you.

AI agents are starting to do more than answer shopping questions. Soon they will decide which products people see and, in some cases, buy on their behalf.

Big marketplaces are building the easiest version of this: an agent that searches one catalog and steers the buyer toward that platform's checkout. That may be convenient, but it is not independent advice. The marketplace still decides what can be seen and makes money when the agent closes the sale.

NorthCinder takes a different approach. It compares products from the sources you choose and shows where its facts came from. Before it buys anything, it asks for your approval. NorthCinder is software you run alongside your AI app.

The repository owner does not operate a NorthCinder service. There is no NorthCinder account or cloud service.

Get started

You need Node.js 20 or later and an MCP-capable AI app.

npx northcinder init

The command saves your configuration on your computer and prints the MCP entry for your AI app. Local mode is keyless. It runs the MCP server and search engine together in one process, using a temporary loopback port.

Once connected, try a real shopping brief:

Find black wool running shoes under $130. Compare price, delivery, fit, and merchant trust. Tell me why the winner ranked first and which options were ruled out.

What a result looks like

NorthCinder does not pretend there is one universal "best" product. It normally shows no more than three useful choices: the strongest fit, a lower-risk option, and a cheaper or meaningfully different option when one exists.

Each result explains why it ranked where it did. You can also inspect the other finalists, rejected offers, and facts that could not be verified.

Research before recommendation

Product research gets messy quickly. Model names overlap, sellers copy one another, and a polished product page can hide the one detail that makes an item wrong for the buyer.

NorthCinder includes separate research guides for products and sellers. Before doing the research, the MCP host should:

1. Read northcinder://research/product or northcinder://research/seller.
2. Call create_research_plan with the actual request and exact subject.
3. Follow the returned checklist with the research tools it already controls.

If the sources disagree or do not identify the exact product or seller, the result stays provisional. Research can decide whether an offer is ready to compare, but it cannot add ranking points.

No host and model combination is currently qualified for routine research use. Treat every research result as provisional until the buyer checks its identity, sources, conflicts, and unknowns.

Buying stays a separate decision

A recommendation is not permission to buy. Every checkout needs a fresh approval for one exact offer and one unit. The signed approval includes the merchant, variant, price, known total, and spending cap. It can be used once.

NorthCinder rejects raw card details. A supported automated checkout can use an opaque payment token, or NorthCinder can hand the buyer a cart link to finish in their own browser.

Order outcomes stay local and only attach to the purchase they belong to. NorthCinder does not silently rewrite the buyer's profile, and its reminders only send notifications.

Rules you can inspect

Seller payment never improves ranking. Sponsored offers stay labeled and below organic results. Missing store coverage stays visible, and unknown seller history remains unknown instead of being guessed safe or unsafe.

NorthCinder reruns the ranking locally and writes recommendations, approvals, and checkout attempts to a local audit log. The ranking specification, trust specification, neutrality audit, and checkout package contain the details.

---

GitHub에서 전체 내용 보기