Chennai, India /
← All work

RepoLens

Project
Repository Intelligence Platform
Focus
Repository intelligence
Platform
Next.js / FastAPI
Status
Independent project
RepoLens repository analysis overview

Completed analysis of Asmit-Rai-04/RepoLens: 124 source files, 184 graph nodes and 605 edges from the real repository.

(Overview)

A repository intelligence platform that turns unfamiliar source code into an explorable structural model.

Reading a large codebase file by file makes it hard to see module boundaries, dependency cycles and structural hotspots.

Technologies
  • Next.js
  • TypeScript
  • Python
  • FastAPI
  • Tree-sitter
  • NetworkX

(The build)

Repository Intelligence Platform

RepoLens securely ingests public repositories, discovers source files and uses Tree-sitter to parse supported languages. NetworkX graph analysis surfaces cycles, strongly connected components and module importance. The frontend connects high-level architecture evidence to individual files.

RepoLens architecture graph
Detected frontend/backend separation and layer dependency flow from the same repository analysis.
RepoLens dependency graph with a selected file’s incoming and outgoing dependencies
Interactive dependency inspection: app/schemas/source.py has 32 incoming edges and one outgoing edge in this analysis.
RepoLens source inspector showing parsed symbols and actual source code
Source inspector reached from the selected graph node, with six parsed symbols and the repository file contents.

(Engineering details)

04
01

Static source analysis without executing repository code

02

Dependency graphs, cycles and strongly connected components

03

Evidence-backed architecture and module drilldown

04

Interactive dependency visualization with display limits

(Architecture)

  1. 01

    Public GitHub URL

  2. 02

    Safe archive ingestion

  3. 03

    Source discovery

  4. 04

    Tree-sitter parsing

  5. 05

    Dependency graph

  6. 06

    NetworkX analysis

  7. 07

    Architecture evidence

  8. 08

    Interactive exploration

(Evidence & context)

The repository documents a completed development analysis of Microsoft TypeScript with 31,329 source files, alongside Kubernetes with 13,764. These are repository-reported validation runs; the portfolio has not independently rerun them. The 32,000-node figure is omitted because its exact context was not established.

Architecture detection uses deterministic heuristics. The analysis store is in memory, graph rendering limits visible nodes, and resource budgets can produce partial analyses.

Read the repository ↗
Next projectFinancial Risk Prediction Engine ↗