Scraper attempts, transparent

What the extractors actually did, last run.

Every scrape attempt against every tool is appended to a per-tool JSONL log. This page surfaces the most-recent event per tool with outcome, plans found, and any warnings the extractor flagged. Useful when /status says "manual-only" but the extractor is actually running and silently falling back to prior verified data.

Companion to /freshness (when data was last verified) and /status (persisted extractor state). Raw machine-readable feed: /api/scrape-runs.json.

Tools tracked: 37
With attempts: 0
Last run: ok: 0
Stale-preserved: 0
Failures: 0

All tools, most-recent attempt first

Tools that have never been attempted appear at the bottom. A scrape attempt records an event regardless of whether it produced a successful write; this is the audit trail of every cron and manual run.

beehiiv Email marketing

never run
no runs
Buffer Social media management

never run
no runs
Buttondown Email marketing

never run
no runs
Cal.com Scheduling

never run
no runs
Calendly Scheduling

never run
no runs
Close CRM

never run
no runs
Convert A/B testing

never run
no runs
Crisp Live chat

never run
no runs
Customer.io Marketing automation

never run
no runs
Demio Webinar

never run
no runs
Drip Email marketing

never run
no runs
Flodesk Email marketing

never run
no runs
Ghost Content marketing

never run
no runs
Help Scout Help desk

never run
no runs
Hightouch Customer data platform

never run
no runs
Hunter Lead generation

never run
no runs
Intercom Live chat

never run
no runs
Kit Email marketing

never run
no runs
Leadpages Landing pages

never run
no runs
Lemlist Sales engagement

never run
no runs
Mailchimp Email marketing

never run
no runs
MailerLite Email marketing

never run
no runs
Mailgun Transactional email

never run
no runs
Mailtrap Deliverability

never run
no runs
Mixpanel Analytics

never run
no runs
Omnisend Email marketing

never run
no runs
PostHog Analytics

never run
no runs
Postmark Transactional email

never run
no runs
Postscript SMS marketing

never run
no runs
Resend Transactional email

never run
no runs
Substack Email marketing

never run
no runs
Surfer SEO SEO

never run
no runs
Tally Form builders

never run
no runs
Triple Whale Marketing attribution

never run
no runs
Typeform Form builders

never run
no runs
Unbounce Landing pages

never run
no runs
Webflow Landing pages

never run
no runs

How to read this page

Each tool has a per-attempt JSONL log at data/scrape-runs/<slug>.jsonl. Every event records run timestamp, outcome, plans found, warnings, and error message on failure. Six possible outcomes:

ok: Extractor succeeded, merge produced a confident write.
stale-preserved: Extractor ran but the merge layer judged the scraped price low confidence; prior verified data was preserved. Tool ends up with scraper_status=stale.
extractor-threw: Playwright crashed or the extractor raised. A screenshot is saved to data/scraped/ for diagnosis.
validation-failed: Merged data did not pass the canonical Zod schema. The most likely cause is an editorial field shape mismatch.
no-extractor: No module at scripts/scrape/extractors/<slug>.ts. Should never happen for a tool in the spine; if it does, see the dry-run gate.
no-tool-json: Extractor exists but no tool JSON stub. Run npm run seed first.