guasi18587278913/creator-breakout-finder

Creator Breakout Finder is a tool for anyone who studies social media accounts and wants to know which posts performed unusually well, not just which ones had the highest raw numbers. When you are comparing accounts, the real question is which posts stood out relative to that creator's own typical performance. This tool answers that by reading a creator's recent public posts and comparing each one against their own baseline, so you are not comparing a small account against a massive one. You paste a profile URL from one of six supported platforms: Xiaohongshu, Douyin, Bilibili, TikTok, YouTube, or X. The tool fetches recent public posts using the user's own TikHub API key. For most platforms it looks at likes, and for Bilibili and YouTube it looks at view counts. It estimates what normal performance looks like for that specific account, then flags posts that clearly exceed that baseline. A post must meet three conditions to be flagged as a breakout: its metric must be at least double the account's historical median, it must rank in the top 10 percent of the sampled posts, and it must hit a minimum threshold of 500 views or 20 likes. Confidence is reported as low, medium, or high depending on how many posts were available. The tool exports a Markdown research report and a CSV file with all the sample data. It also ships with fully offline demo data, so you can try the entire workflow without a key. The tool deliberately rejects individual post links, short links, non-HTTPS URLs, and URLs with embedded credentials or custom ports. API keys are read only from a local environment variable, never exposed in browser forms, error messages, or exported reports. Raw API responses are kept in memory only and never written to disk. Anyone deploying this publicly must add their own authentication and rate limiting. The README is honest about limitations. The analysis is a one-time historical snapshot, not a real-time growth curve. Older posts have had more time to accumulate views and likes, so results are best used as a starting point for deeper research, not as proof of why something went viral.