Trace

Overview

Trace solves a painful problem in modern codebases: understanding how requests actually move across services, routes, clients, and downstream dependencies. In most teams, architecture diagrams are outdated and onboarding depends on tribal knowledge. Trace replaces that with source-of-truth mapping directly from code. The platform has two tightly integrated parts: a local CLI scanner and a web dashboard. The CLI parses repositories using AST analysis, detects routes, HTTP calls, GraphQL operations, and key references, then builds a structured relationship graph. That graph is uploaded to the platform, where teams can explore it as an interactive DAG and inspect API paths end to end. From an architectural perspective, Trace uses a monorepo setup with a Node.js CLI and a Next.js App Router frontend backed by PostgreSQL via Prisma. Visualization is powered by React Flow with Dagre layouting, enabling readable graph rendering at scale. Authentication and team-level project workflows support secure collaboration and API-key-driven ingestion from local machines or CI pipelines. A core challenge was balancing deep static analysis with practical speed and reliability across different project structures. To address this, the scanner combines parser-driven extraction with resilient fallback logic, configurable include/exclude patterns, and depth controls for large repositories. The result is a fast, developer-friendly workflow where one command can turn a complex codebase into a navigable architecture map in seconds.