Audit 38 fresh Show HN launches for AI-assistant readiness: eight page checks plus asking a model if it knows the product
A one-person SaaS founder pulled a month of Show HN launches through the Hacker News API, ran eight technical checks on the top 38 sites and asked one language model whether it knew each product. It could name 3.
Evidence: The source shows its work: steps, screenshots, code or data you can inspect.
The business problem
A founder launching a small product wants to know whether AI assistants can find and describe it, and which basic page signals are usually missing at launch.
What was tried
1. Pull recent Show HN stories through the free Hacker News Algolia API (400 stories over 30 days, 227 unique product domains after removing platforms). 2. Take the top 40 by score; drop any that return nothing readable (38 left). 3. For each site, run eight checks with a single HTTP request each: robots.txt, sitemap.xml, llms.txt, and on the home page the title, meta description, JSON-LD and a single h1. 4. Ask one model (deepseek-v4-flash, one provider, one day) whether it knows the product and what it is. 5. Tally failure rates per check and compare pass rates with Hacker News points. A founder can run the same eight checks and the same question on their own site before launch.
What was reported (not reported)
The model could not name 35 of 38 products. Common gaps: broken or missing title or description (61 percent), no JSON-LD (53 percent), no llms.txt (47 percent), missing sitemap (32 percent), no single h1 (29 percent). No site passed all eight checks; the average passed five. Hacker News points had a weak negative link with pass rate (r = -0.21). The author's own site scored 7 of 8 and was still not named. No result is reported for whether fixing the checks makes a model name a product.
Limitations
Vendor: the author runs SaaSClerk and links to its free AI-visibility scanner. One model, one prompt, one day; other assistants were not tested. No JavaScript rendering, so client-rendered sites may be misread. Show HN is a small, technical sample. Being named is a minimum and says nothing about being recommended. The checks are correlational, not shown to cause a model to know a product.
What you need
Ability to call the Hacker News Algolia API and a model API, or to run the checks by hand on your own site.
Sources
- DEV Community (fuxiaohua, SaaSClerk) ↗ Vendor case study, published October 30, 2026
Source published: October 30, 2026. Last reviewed here: November 1, 2026. Spot a mistake? Tell us.
Tools in this workflow
- Hacker News Search API (Algolia): Source of recent Show HN launchesTry Hacker News Search API (Algolia)
Get new workflows like this by email
A short selection of the most useful new workflows, including the ones that did not work. Free, and you can unsubscribe at any time.
Related workflows
Rank what to build next with an open-source MCP server that scores support tickets and feature requests, then lets Claude write the plan
A product lead built an MIT-licensed MCP server that pulls Help Scout tickets and ProductLift requests, strips personal data, groups them into themes, scores them with a weighted formula and hands the result to Claude, with business KPIs able to reorder the list.
When an AI marketing agent's test and your own test change the same clips on the same day
In a public log of AI agents selling small products, a marketing agent proposed a sound test of why TikTok likes fell. The operator found the data could not answer it, because their own music test had started on the same day on the same clips.
Sixteen weekly AI reviews that raise tasks with a predicted metric and open pull requests
A mostly solo SaaS founder runs one weekly AI review per area of the business. Each review writes a report, turns problems into tasks that predict which number should move, ships approved fixes as pull requests, and scores them the following week.
A scheduled Hermes Agent that watches a marketplace side business and acts on what it finds
A seller of 13 web-scraping tools on the Apify Store runs a Hermes Agent on cron each morning: it scans platform news, checks pricing compliance and listing gaps, prepares fixes, and sends an 8-line report.