Comet compares coffee makers and prepares a cart
Brian Glassman’s DreamHost test reports Comet comparing coffee makers under $100, selecting two for an Amazon cart and stopping before purchase.
dreamhost.com
Original prompt
Shared by the source author. See source notes for context.
Go to Amazon.com, research the best coffee makers under $100 with coffee bean grinding instead of pods,
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- Brian Glassman
- Original account
- Read the source ↗
- Published
- 5 Jan 2026
- Source link checked
- 15 Oct 2026
- Reported result
- The DreamHost testers report five compared candidates and two items added to the cart, with no checkout attempted.
- Human involvement
- The testers set the shopping constraints and inspected the cart endpoint; purchase was outside the task and not completed.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Ask for five coffee-maker comparisons within the price and feature limits.
- Let Comet inspect Amazon results and assemble the comparison table.
- Add two selected candidates to the cart and stop before purchase.
DreamHost tested several browsers on the same shopping task. Brian Glassman's account asked for coffee makers under $100 with bean-grinding rather than pod use, a comparison table of five choices and two candidates placed in an Amazon cart with the supplied delivery location.
The Comet run searched the retailer's results, identified options matching the requested price and features and produced the table. The testers report that it added two candidates to the cart and stopped there. No purchase or delivery is documented.
The distinction matters because a useful shopping result combines research with a reviewable state in the store. A recommendation table alone would leave the shopper to repeat the search; an unapproved checkout would go beyond the stated task. Here, the reported endpoint was the cart page.
The article compares multiple browsers, but this entry attributes only the described Comet result. The products, prices and model numbers are not independently verified or treated as current recommendations. The test does not establish that every selected coffee maker truly met all requirements or that Comet works reliably on every retailer. The evidence supports one reported end-to-end comparison and cart-preparation trial under a specific brief.
Prompt
Original prompt
Source: Blog · Brian Glassman
Frequently asked questions
Was anything purchased?
No. The source says Comet stopped at the cart page without completing a purchase.
Which requirements were tested?
A price under $100, coffee bean grinding rather than pods, five comparison candidates and two cart selections.
Are the listed prices current recommendations?
No. They belong to a historical test and were not independently rechecked for this entry.
More details
About this account
Comet searches Amazon, compares five products and adds two choices to the cart.
Firsthand primary account opened and read; original publication date checked in primary metadata or visible timestamp. Outcomes are author-reported, not independently reproduced. Historical comparison test; no purchase, current price validation or independent feature audit. Original prompt excerpt reproduced exactly from the source; excerpt only, not the complete prompt.
Credit: Brian Glassman
Source published 5 Jan 2026
Source checked 15 Oct 2026
Tools and topics
Perplexity Comet · Amazon · Shopping cart
Shopping · Reported deployment
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.