Understand any codebase.
Turn a GitHub repository into a navigable map of dependencies, symbols, and architectural risks.
True semantic AST parsing for symbols, functions, types, and exports.
Strongly connected components isolation mathematically finds recursion.
Every file and graph node bound to an immutable GitHub commit revision.
Static analysis without running untrusted code or saving data to disk.
Built for engineering teams who demand precision
Everything inside VEYRO is backed by genuine parsing pipelines and verified algorithmic graph topologies.
Real AST Parsers, Not Regex
Uses the official TypeScript Compiler API and Python AST parser to extract genuine symbols, type interfaces, and import declarations. Never fooled by comments or multi-line strings.
Tarjan SCC Cycle Detection
Finds circular dependency loops mathematically using Tarjan’s strongly connected components algorithm. Highlights affected files so you avoid nasty runtime undefined exports.
Deep Import Resolution
Resolves relative specifiers, directory index files, external package dependencies, and Python standard libraries. Broken or unresolvable imports are transparently reported.
In-Context Code Viewing
Embedded CodeMirror 6 client viewer with syntax highlighting and line numbers. Selecting any graph node or tree leaf pulls the raw file content at the pinned commit revision.
Global Command Palette (Ctrl+K)
Instant keyboard-driven search indexing file paths, symbol declarations, external package imports, and architectural findings across the entire analyzed repository.
Honest Metrics & Denominators
Parser coverage calculates an explicit denominator (analyzed / discovered files). No made-up scores or fake 100% badges—real engineering diagnostics.
How VEYRO analyzes code without cloning
A strictly typed, bounded pipeline designed to inspect remote repositories with speed, security, and algorithmic precision.
GitHub Ingestion
Resolves repository to an immutable 40-character commit SHA. Recursively fetches tree metadata and filters out vendor bundles (node_modules, dist, lockfiles).
Semantic Parser Registry
Routes files to language-specific adapters. Generates semantic symbol tables, import declarations, and structural exports without resorting to regex guesswork.
Resolution & Topology
Maps specifiers to target file nodes. Computes incoming and outgoing degrees, identifies architectural hubs, and executes Tarjan’s algorithm for cycle detection.
Explorer Workspace
Renders the deterministic dependency graph using Dagre layout. Synchronizes file selection, source inspection, and findings via a reactive client store.
Language support & runtime limitations
We don’t promise fake magic. Here is exactly what VEYRO supports today and the engineering boundaries we enforce.
Rate-Limit Transparency
Unauthenticated GitHub API calls are subject to GitHub’s 60 req/hr IP limit. If exhausted, VEYRO provides clean reset diagnostics and HTTP 429 notifications.
Max Analyzed Files (200 Cap)
To prevent browser lockups on massive monorepos, VEYRO bounds AST parsing to the top 200 relevant code files, prioritizing source code over build outputs.
500KB Max File Size Cap
Minified bundles or large binaries exceeding 500KB are skipped from AST parsing to preserve memory and maintain responsive UI rendering.
Bounded Concurrency (5)
Concurrent file fetches are clamped to a pool of 5 simultaneous HTTP requests to stay within network thresholds and avoid secondary rate limits.