Gevalinho Labs

**CodeArchaeologist AI** is an AI-powered repository intelligence platform designed to help developers understand unfamiliar, undocumented, and legacy codebases. Developers often spend days or weeks manually investigating software architecture, dependencies, hidden business logic, and application workflows before they can confidently maintain or modernize existing systems. CodeArchaeologist AI addresses this challenge through automated GitHub repository scanning, isolated source-code analysis, architecture mapping, workflow investigation, and evidence-grounded AI reasoning. **Powered by Meta Muse Spark**, accessed through **Baseten's AI inference infrastructure**, the platform enables developers to investigate repository architecture, identify components and relationships, explore candidate business rules, and ask natural-language questions about source code. Unlike generic AI code summaries, CodeArchaeologist AI connects findings to source-file and line-number references, distinguishes observed facts from inferred behavior, and explicitly acknowledges incomplete evidence rather than generating unsupported conclusions. **Technology and Infrastructure:** - **AI Model:** Meta Muse Spark (`meta-models/muse-spark-open-1026`) - **AI Inference:** Baseten - **Application Hosting:** Amazon Web Services (AWS), using an Ubuntu server, Nginx, and isolated Docker-based scanning - **Application Framework:** Next.js, React, and TypeScript - **Source Integration:** GitHub The platform is deployed and publicly accessible at: **https://codearchaeologist.theconduitbox.com** Public source code: **https://github.com/gevalinho/codearchaeologist-ai** CodeArchaeologist AI aims to accelerate developer onboarding, legacy software investigation, technical documentation, and modernization planning by transforming complex repositories into understandable, evidence-backed architectural knowledge.