Guide

How to find a tool, read the numbers, and follow the lists.

Describe what the tool should do in a short sentence, such as “PDF to Markdown with OCR”. Product names work too.

Search works best in English. A search in another language is translated into English first, and the results page shows the English query it used.

Results come closest first. Those related by meaning but less close are listed under “Looser matches, by meaning”.

A query longer than 200 characters is shortened to its first 200.

Browse Trending and Latest

Trending lists the products that drew the most reaction in the last 3, 7, or 30 days (7 by default). Latest lists every product, most recent reaction first.

Data is collected once a day. The line under the Trending heading shows when.

Read the numbers

Each source has its own column, and each number is that source’s own:

HN Hacker News
points on the Show HN or Launch HN post
PH Product Hunt
votes on the Product Hunt launch
GN GeekNews
points on the Show GN post
GitHub GitHub
stars on the GitHub repository

If a product has several posts on one source, the column shows the highest number. On Trending, it shows the post that counted for the ranking. An empty cell means no reaction from that source, and “–” means a reaction without a number.

“3 reactions” counts every post about the product. “Latest reaction” is when the most recent one was posted.

How Trending ranks

Each post is compared with the same source’s posts from the last 14 days, as the share of them it beats, from 0 to 1. A product’s value is the best of these from each source in the period, added together.

Numbers from different sources are never added as they are: 100 Hacker News points and 100 Product Hunt votes mean different things. Turning each number into its place within its own source makes them comparable.

For GitHub, the stars a repository gained in the period are compared with its pace over the 28 days before. Repositories that grew faster than usual are taken fastest first, as many as the average number of Hacker News and Product Hunt posts in the period; the first gets 1 and each next one a little less. GeekNews points are shown on each product but are not counted in Trending yet.

Read a product page

“Reactions” lists every post about the product with its source, number, comment count, and date, linked to the original discussion.

“Numbers and dates by source” shows the highest number from each source and the dates of the first and latest reaction.

“Similar products” are products with the closest descriptions.

“Link not responding” means the product’s site did not answer when last checked, with the date of the check. “Archived” means its GitHub repository is archived.

Follow a list

Trending is also published as an RSS feed and a JSON Feed. The links are under the Trending heading and follow the period you picked. Daily digest sends one entry a day with the top 5.

Feed addresses take these parameters:

period
3d, 7d, or 30d
q
search words, to narrow the list
forms
CLI, Library, Desktop, Mobile, Browser extension, Self-hosted, comma-separated (a product must have all of them)
open_source
true or false
n
number of items, up to 50 (20 by default, 5 for the digest)
format
digest for one entry a day

Example: https://tangerineindex.com/en/trending/feed.xml?period=30d&q=pdf

Ask from an AI assistant

Claude, ChatGPT, Cursor, and other MCP clients can search the same products and Trending. Set up MCP