AI Tool Discovery Agent
For teams who ship against new tools
Find the AI products that launched this week. Leave the roundups in the search results.
The agent plans a date window, searches SerpApi, compares web and news, and posts a card to the console, Slack, or Microsoft Teams.
Ready to launch discovery sequence.
Click "Start Pipeline" to trigger discovery run.Who it is for
A team lead who needs to know what shipped, not what a blog ranked. A researcher who wants the product page, not the article about the product page.
The useful part is the rejection. Search returns launch lists, museum pages, and news. The digest keeps a product: what it does, who it is for, what it costs, and where it lives.
01
A date window, not a hope
Google web search takes SerpApi’s tbs=qdr:dN. Google News takes after:YYYY-MM-DD. Both stay inside the days you pass.
02
Two engines, one list
Web and news run together. The same URL collapses to one row. When the dates disagree, the news date wins.
03
Articles point at products
A launch post that names one product homepage is fetched next. The digest keeps that homepage. A listicle that names several products is dropped.
04
A card, not a paragraph
Each kept page becomes title, audience, summary, features, pricing, and source. The console prints it. Slack or Teams can post it.
The same card, two rooms
Illustrative output: Teams on one side, Slack on the other. A phone shows both, one after the other.
Teams
SlackWhat happens when a step fails
- One engine fails.The other engine’s rows are kept. The run continues.
- Both engines fail.The process exits with an error. It does not print an empty success.
- A page returns 403, or robots.txt says no.That URL is skipped. The rest of the batch continues.
- The title is already a list.“10 new AI tools” is dropped before fetch. One named product site is fetched next.
- Slack or Teams is unset.That delivery is skipped. The console still prints the cards.
- Dry run.Search and compare only. No fetch, no extraction, no post.