Understanding results
Sources
When a platform browses the web before answering, it cites the pages it drew on. Pineprompt records every one of those citations and organizes them on three levels: the citation itself, the URL it points at, and the source, which is the domain the URL belongs to.
The three levels
- Citation
- A single reference inside a single response: the URL, its title, and its rank among the response's citations.
- Source URL
- A specific page, aggregated across every response that cited it. This level answers which exact article or listing shapes the answers.
- Source
- The domain, aggregated across all of its URLs. This level answers which publications and platforms the platforms trust for your prompts.
Both aggregated levels carry the usage metrics described on the citations page: the percentage of responses citing them, the absolute count, and the average number of citations per citing response.
Source types
Every source is classified by the kind of site it is: news, blog, forum, social, video, review, directory, webshop, business, wiki, documentation, academic, government, or institutional. The classification lets you read the citation mix at a glance. Platforms answering your prompts from forums and review sites are relaying community opinion; platforms citing documentation and news are relaying publishers. The two situations reward different work.
Reading the source views
The project overview lists the top sources and top URLs by usage under the active filter. The sources section of the app holds the full ranked list, and each source has a detail page showing its URLs, its usage over time, and the responses that cited it.
Typically, the most useful reading is comparative: filter to a tag or a platform where your visibility disappoints, and look at which sources dominate there. Those pages are the input the platforms are working from, and earning an accurate presence on them is the most direct lever you have on what the answers say.
A note on platforms without citations
Not every platform cites sources on every answer; some answer from their training data without browsing. A response without citations is normal and still counts fully toward the mention-based metrics. Source usage percentages are computed over all responses in the filter, so a platform that rarely cites will simply show lower usage figures across the board.