valeriyalexeev/Experteasing

What it is for Experteasing assembles a short panel of real, web-sourced specialists for a given topic and country, runs a structured discussion (debate, council, Delphi, red team, etc.), fact-checks decision-relevant claims, and stores consolidated memory so the council "remembers" and improves across sessions. Who uses it Decision-makers, product teams, founders, researchers or anyone needing structured expert input and reproducible rationale for decisions. Key features - Finds real specialists (not fixed personas) and enforces genuine disagreement. - Web-based fact-checking of sharp claims with sources. - Compounding, human-readable memory using an LLM-Wiki pattern (local markdown index + per-expert wiki). - Multiple discussion formats and output depths (one-pager, transcript, highlights). - Optional arbiter that scores arguments. - Runs locally via Claude Code (requires Claude subscription and web search); no API keys or servers; all files stored on your machine.

AI-related keywords found in this repo's description, topics, or README summary — grouped by category. Each badge links to the corresponding ranking detail page.