HALDxAI: evidence-routed knowledge graph software for human aging and longevity research
Abstract
HALDxAI is a knowledge graph platform for exploring human aging and longevity through literature-derived relations, integrated biomedical resources, and source-linked evidence. This versioned archive contains the previously released source code, an updated backend snapshot, and selected validation materials accompanying the manuscript “A systems-level dissection of aging and longevity mechanisms powered by an evidence-routed AI knowledge graph.” The archive includes:• The original v1.0.0 software source, with configuration files, prompts, and environment specifications.• An updated backend snapshot, including relationship retrieval and agent tools.• Validation scripts, selected de-identified inputs, relationship-control pair lists, and reference results.• Instructions for reproducing the included statistical tables and file checksums. The original software is preserved separately from the updated backend. Historical question-answering results describe the evaluation-time implementation and should not be interpreted as performance measurements of the updated backend. The archive does not include a tested release of the current frontend or duplicate the large database files. Data archive: https://doi.org/10.5281/zenodo.18812914Development repository: https://github.com/ShilongZhang116/haldxai HALDxAI supports evidence exploration and hypothesis generation. Retrieved relations and graph paths require interpretation in the context of their supporting sources.
