MCP on Hacker News, month by month, with the sparkline drawn in the query
One statement over hackernews.items: monthly counts of the token ‘mcp’ since January 2025, with bar() drawing the sparkline in the result set. A hump, not a line: the peak was March 2026. hasToken is a whole-token match over a full-text index, so it counts the token, not the sense; rows land every few minutes, so your run will differ from this one by the day. Measured 2026-09-09: 0.76 s, 6,027,149 rows read, $0.0030 of machine time, $0 charged.
Evidence
20 rows 6,027,149 rows read 765 ms complete machine cost $0.003009 charged $0.00 · free slack
| m Date | n UInt64 | spark String |
|---|---|---|
Run it yourself, or change a word: sign in · create an account.
Query
SELECT toStartOfMonth(original_timestamp) AS m, count() AS n,
bar(n, 0, 2200, 20) AS spark
FROM hackernews.items
WHERE original_timestamp >= '2025-01-01'
AND original_timestamp < toStartOfMonth(today())
AND hasToken(search_text_lc, 'mcp')
GROUP BY m ORDER BY m LIMIT 100Continue this research
This page is legible to agents — claim, rows, SQL, and provenance are one request away in plain markdown.
Your agent
Hand any agent this share and Scry's self-describing surface.
curl https://scry.io/s/dikjdnv0?format=md https://scry.io/llms.txt — schema, query API, and rulesThis page
Take the whole finding as markdown for a prompt, notebook, or doc. https://scry.io/s/dikjdnv0?format=md