I built a cross-platform Discover endpoint so AI agents can answer "what are people saying about X" without wiring five scrapers. Looking for feedback.

I've been working on a new API endpoint called "Discover" for agent builders. The idea came from my customers (I'm building SocialCrawl). They wanted their agents to find out 'what people are saying about X' across different social platforms at once. But they wanted to do that without wiring a handful of scrapers for each platform and filtering out their data of interest like demographics, gender, etc.. So I built an API endpoint for that. Topic words go in query, the actual question goes in context. It searches several platforms (reddit, linkedin, tiktok you choose), keeps a date window, judges every row against your question, removes duplicates, and returns one list. Example looks something like this: (depth=quick, period=90d): query=electric cars context=What do people in the UK think about owning an electric car? A few details: Context judging: Rows are scored against your natural-language question, not just a keyword match. Small or large: Small requests answer in the call. Bigger ones return 202 and continue as a run you can follow. Spend controls: dry_run=true prices a call for free. max_credits caps spend and never overspends. Other views: Same call can return posts, accounts, communities, or items (hashtags, sounds, named things). It's live, and I'd like any feedback you might have! Also giving out free extra API credits if anyone wants to test this properly - let me know. More info in the comments!