producthunt.com
top 3 products of the day on the homepage: product name, tagline, and upvote count
- Collector
- c_mt5g0w2v22o0ro7nzf
- Last run
- 2h ago
- Rows
- 3
- Repairs
- 2
Rows per run
21 runs recorded1 came back wrong
Fields under watch
The shape of the last good run. If one of these stops coming back, that is what triggers a repair — and what gets described in the repair prompt.
Repair history
- Repaired and verified
Scope the scraper to the top 3 ranked products only (was returning all 28), fix the duplicated tagline text, and require upvote_count on every returned row.
- Repaired and verified
The scraper is returning bad data. Problems found: Schema drift: missing upvote_count; Field "upvote_count" is empty in 100% of records. The page layout has most likely changed and the selectors need updating.
The data it collected
Every healthy run overwrites data/producthunt.com/latest.json in the repo and keeps a timestamped copy beside it. The file below is public: no key, no API, nothing to sign into.
3 rows · 4 fieldslatest.json ↗
| product_name | tagline | upvote_count | product_page_url |
|---|---|---|---|
| Construct Computer | Your AI coworker gets a computer. You get your day back. | 13 | producthunt.com/products/construct-computer |
| FetchSandbox | API integration testing that remembers what breaks | 9 | producthunt.com/products/fetchsandbox |
| Aximote | Your car data, finally in your pocket | 7 | producthunt.com/products/aximote |
source · https://www.producthunt.com
Running and repairing this scraper on demand happens through your coding agent today — scraper_run and scraper_heal. Buttons here are the next piece of work, not a shipped feature.