— LEAD STORY · RANKED BY THE NEWSROOM ENGINE —
THE HEALING LAB
Break a site. Watch Spider-Sense notice. Watch bdata scraper heal put it back. Nobody writes a selector.
STEP 1 · PICK A VICTIM
Nothing broken yet. Every collector is in print.
HEALTH
—
RECORDS
—
FIELDS
—
STEP 2 · THE REPAIR, LIVE
awaiting orders. break a collector, then heal it.
idle
0.0s
THE COLLECTOR FLEET
Eight niche developer tools with no pre-built scraper, no API and no RSS. Each one is a Scraper Studio collector generated from one sentence of plain English.
OFF THE WIRE
Every changelog bullet, scored for what it will actually cost you. Breaking changes and pricing moves lead; chores get buried on page nine.
THE CANON
Every break, detection, repair and recovery, in order. This is the audit trail that makes “self-healing” a claim you can check.
HOW THE PRESS RUNS
One sentence in, a production API out, and a repair loop that closes itself.
1
Describe
One sentence of plain English per target in data/targets.json.
▸
2
Generate
bdata scraper create compiles it into a collector. You own the code.
▸
3
Trigger
POST /dca/trigger — a production endpoint with no infra of ours.
▸
4
Sense
Four content signals score the payload. Silent failures get caught.
▸
5
Heal
bdata scraper heal with a brief derived from the failure itself.
▸
6
Verify
Re-scrape, re-score. Only a passing payload closes the loop.