Niches · Shopping, returns & refunds

Buy for the shopper after they approve

No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.

0companies, full fit
3companies, fit with a limit
31companies carry the tag
20share of the category's work
7first-hand accounts

What counts as done

An order is placed through an API or MCP tool only after an explicit confirm step showing merchant, items and total, the charge cannot exceed the approved amount, an order id comes back, and a retried call does not place a second order.

How agents ask for it: buy this for me, check out with my saved card, order it if it's under $50

Does the job, with a limit 3

It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.

ListingWhy
MealMe API E
MealMe, Inc.
Create Order with place_order false returns the final quote, then true places it, and charge_user bills total_with_tip to the saved payment method. No idempotency key; customer_order_id only helps look an order up before a retry. source
opencommerce C
Agentcard Corporation
confirm is set only after the shopper's yes, the passkey approval page shows store and amount, the hold ceiling is max_charge, and order_id comes back. No idempotency key: after a timeout the agent reads the conversation before resending, and store terms on automated buyers go unaddressed. source
Shop MCP B
Shopify Inc.
complete_checkout runs only from ready_for_complete after the shopper confirms totals, with meta.idempotency-key, and spend requests carry a limit. Sign-in needs a registered client or a client ID metadata document with no dynamic registration; acc-23 and acc-24 hit that. source

Couldn't check 2

The docs were closed to us, or silent. These carry no mark either way.

ListingWhy
Best Buy Developer API E
BestBuy.com, LLC
The public APIs cannot order. The Commerce API that places orders is invite-only, with documentation shared after approval, so its confirm step and retry rules could not be read. source
Walmart Agent Gateway F
Walmart Inc.
The post says customers approve agent purchases by default or set a spending limit in Agent Profiles, launching within a month, and agents never see card data. No order tool or retry rule is published. source

Doesn't do this job 7

Listings in the category that do something else, with the reason.

ListingWhy
Amazon Creators API F
Amazon.com Services LLC
No order or checkout operation; the operations are SearchItems, GetItems, GetVariations and GetBrowseNodes. source
Etsy Open API v3 B
Etsy, Inc.
No buyer checkout or order creation endpoint; receipts are the seller's, behind OAuth. source
Instacart Developer Platform API + MCP E
Maplebear Inc. dba Instacart
No order call; checkout happens on Instacart after the hand-off link. source
Ocado F
Ocado Retail Limited
Orders are placed on the website or app by the customer; no ordering route for an agent is documented. source
Too Good To Go F
Too Good To Go ApS
acc-27 could reserve stock through the unofficial API but not pay or complete the order outside the app; no official order route exists. source
Walmart Affiliate Marketing API F
Wal-Mart.com USA, LLC
Read-only catalogue API; the operations are lookup, catalogue, reviews, search, stores, trending, taxonomy and recommendations. source
Zillow in Claude F
Zillow Group, Inc.
All nine tools are read-only; nothing is bought. source

What stops agents finishing

  • retailer terms bar buy-for-me agents (eBay, Amazon against Perplexity)
  • OAuth sign-in rejects the agent client
  • no idempotency key on order creation
  • bot check on the only route

Around it

Before: shopping.find-compare, shopping.list-to-cart, human-in-the-loop.approve. After: shopping.order-status.

Where this comes from

Weight 20 of 100 in Shopping, returns & refunds: 7 accounts (acc-01 to acc-03, acc-10, acc-23, acc-24, acc-27); most are about trust or sign-in, few are completed purchases

The same as data: /api/v1/use-cases/shopping.place-order.json, or the MCP tool get_use_case with id shopping.place-order. Other jobs in this category: /api/v1/niches?category=shopping.

Draft, read 2026-11-15 from public docs and first-hand accounts, not yet reviewed by a person. A mark says what the docs show, not what a run measured.

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.