Quietly sharing something I’ve figured out: a next-level use for Astra.
One thing up front: this is about public sites that don’t require a login. It’s a different situation from what I wrote before about not brute-forcing login-gated sites, so the two don’t conflict.
For sites with thorough anti-scraping but too niche for any third-party scraping company to cover:
- Send Astra to explore the site first, using its strong browser-control skills to try small batches.
- Write the process into a detailed technical doc/SKILL, meant for a weaker model to use later.
- Hand it off to Luna Max plus a goal, and let it set its own random batch sizes and random intervals, driving your browser accordingly. Time isn’t the point.
I’ve used this method to pull over three thousand pages off a forum in a week, no trouble so far (I’ll report back if something goes wrong). Burned less than 30% of the $20 weekly quota.