alphasift

Automate A-share stock screening by using strategy YAML files for L1 hard filtering across the entire market, then applying LLM ranking to sort candidates and output structured results for downstream AI analysis.

Yes, you can filter the entire CN stock market by defining custom hard conditions in a strategy YAML file, which the engine applies as an L1 hard filter to eliminate non-compliant stocks from the candidate pool.

LLM ranking compares stocks within the filtered candidate pool by evaluating context, style, and risk to generate an auditable ranking, which is then optionally refined by L3 post-analysis using daily K and industry heat anchors.

Evaluate stock screening strategy performance by saving your execution runs, which enables T+N post-evaluation to track how selected candidates perform over subsequent trading days.

You need external market data including news, announcements, capital flows, and industry heat anchors injected as candidate context to improve LLM sorting stability and ensure accurate risk-adjusted rankings.

Yes, the engine supports pluggable L3 post-analysis, allowing you to integrate external analysis services or local scorecards for deep evaluation and risk adjustment of the ranked A-share candidates.