quant-mind

Why we included this project

Turning research papers, news wires, and filings into something a model can actually reason over eats a large part of a quant researcher's week. QuantMind does that step for you: it ingests raw sources and emits typed knowledge cards that keep their citations and an as-of timestamp, so downstream retrieval gets a traceable, repeatable foundation instead of ad-hoc scraped text. The framework is deliberately agent-native, meaning you can open the checkout, describe the pipeline you want, and a coding agent assembles it against the repo's contracts and deterministic verification. It also works as a normal importable Python library if you prefer to hand-wire the flows yourself. Teams doing LLM-assisted fundamental or quantitative research who need traceable extraction and retrieval rather than one-off scripts will find a coherent structure to build on.