All notable changes to this project will be documented in this file.
- Added
footnote and aiDisclaimer fields to video metadata outputs indicating AI scoring.
- Added native B2B shop contact extraction (
shopName, shopUrl, shopEmail, shopPhone, shopAddress, isShopCreator) for TikTok Shop profiles.
- Added local regex-based keyword filtering for username searches (
Keyword By Username).
- Added robust URL normalizer resolving short formats (like mobile redirects) and sanitizing query tracker parameters.
- Fixed user verification boolean checks to strictly parse boolean states (
=== true).
- Fixed empty text field issues by implementing multi-stage fallback structures (desc, title, caption, AI suggested angle).
- Fixed pagination limits and scraper failure loops by implementing automatic self-healing delegation to a secondary fallback actor.
- Initial release of TrendPilot TikTok Scraper.
- Implemented outcome-specific B2B runModes.
- Added proprietary scoring models: Virality, Creator Authority, Brand Fit, Lead Score.
- Added legacy-compatible output field preservation fallback mode.
- Integrated optional OpenAI, Anthropic, and Gemini LLM enrichment.