Hi HN,I built MyNewsInsights, a political intelligence/research platform for Malaysia.It indexes 274,390 Malaysian political news articles from ~25 years of coverage and turns them into:- entity profiles for politicians, parties, orgs, and public figures - searchable quotes, contradictions, timelines, and co-mention graphs - LLM-extracted financial flows: contracts, donations, fines, subsidies, settlements - Sankey + network visualizations - AI Studio for citation-backed briefings, research chat, and experimental political simulationsThe stack is React + TanStack, Go/Fiber, Postgres, Redis, FastAPI, vector search, D3, and an LLM extraction/validation pipeline.Browsing is free. Login is only needed for AI Studio because of quota/LLM costs.The project is not open source right now, but I’m happy to answer technical questions about the architecture, extraction pipeline, entity resolution, crawler, caching, or the AI agent design.Would love feedback from HN on whether this feels useful for journalism/research/civic tech, and where the trust/citation layer needs to be stronger.Not affiliated with any news outlet.