Legal

Official data sources & evaluation harvesting

How Borsat collects public exchange-site data for evaluation builds — politeness rules, labeling, and how to object.

Evaluation harvest only

Until licensed exchange feeds are active, Borsat may harvest publicly published files and pages from exchange-operated websites only. Every row is labeled data_mode: demo with evaluation attribution and is excluded from commercial entitlements and bulk export.

What we collect

  • Daily trading bulletins → EOD OHLCV and volume
  • Listed securities directories → reference data refresh
  • Announcement titles, dates, categories, links, and ≤2-sentence extracts (never full article bodies)
  • Structured financial figures for Shariah (AAOIFI) inputs from published filings — numbers only, with filing links

Politeness & robots

  • User-Agent: BorsatEvaluationBot/1.0 (+https://borsat.ai/bot; eval-data@borsat.ai)
  • ≥3 seconds between requests per domain; daily caps; backoff on errors
  • robots.txt Disallow is always respected — no override exists
  • CAPTCHA or bot-wall → source marked blocked; we stop

Objection / takedown

Email eval-data@borsat.ai. We disable the source the same day via kill-switch and log the action.

Licensed supersession

When a licensed binding activates for an exchange, scrape-sourced bars are superseded automatically and the scraper for those datasets is disabled (unless kept solely for reconciliation).