Noureddine RAMDI / CodeFlow: Static code analysis and architecture visualization entirely in the browser

Created Tue, 05 May 2026 13:37:39 +0000 Modified Sat, 23 May 2026 20:41:27 +0000

braedonsaunders/codeflow

CodeFlow tackles a problem every developer knows too well: understanding complex codebases quickly and visually. What sets it apart is how it fits full static analysis of dozens of programming languages into a single browser page — no installs, no servers, just your browser working directly on your GitHub repos or local code.

What CodeFlow does and how it works

At its core, CodeFlow is a zero-install, browser-only web application that transforms any GitHub repository or local codebase into an interactive architecture map. It parses over 35 programming languages to extract functions, dependencies, and git churn data. This information is then rendered as a force-directed graph that you can explore with different visualization modes: folder structure, code layers, churn heatmaps, and what they call “blast radius”—a way to analyze the impact of changes across the dependency graph.

The app runs entirely client-side, meaning all parsing, analysis, and rendering happen in your browser without any server-side processing or data collection. For private repositories, it supports GitHub Personal Access Tokens (PAT) to authenticate and increase API request limits.

Under the hood, the tool dives into Abstract Syntax Trees (ASTs) of source files to pull out code elements and their relationships, then combines that with git history to provide insights like code ownership and change frequency. It also performs automatic security scanning, flagging secrets, SQL injection risks, and unsafe eval usage.

Additionally, CodeFlow detects common code patterns and anti-patterns such as singletons, factories, and React hooks. It assigns an A-F health score to the codebase based on these analyses, providing a quick metric to gauge overall architecture quality.

The visualizations are interactive and designed to help developers and architects get a sense of the code structure and risk areas without manually digging through files.

What makes CodeFlow technically interesting

The most impressive aspect is that CodeFlow accomplishes all this heavy static analysis in the browser, within a single HTML file, without any build step or backend. This client-only architecture is rare and demands careful engineering around parsing efficiency, memory use, and API rate limiting.

Supporting 35+ languages means the parser is quite comprehensive. The tradeoff is that it can only analyze syntax and dependencies from source code and git metadata available via the GitHub API or local filesystem, so some dynamic or runtime behaviors are naturally out of scope.

The force-directed graph visualization provides immediate visual feedback, but it can become cluttered with very large codebases. The multiple visualization modes help mitigate this by letting users focus on different aspects of the architecture.

Security scanning and pattern detection run in tandem with dependency analysis, giving a multidimensional view of code health. This integrated approach is valuable but limited to the static signals the tool can extract.

CodeFlow’s GitHub API usage has clear rate limits: 60 requests/hour without a token and up to 5,000 requests/hour with a Personal Access Token or GitHub App installation. This affects how frequently you can analyze large or multiple repositories.

The codebase’s quality is surprisingly clean for a client-side web app of this complexity. The single-file app approach means the entire logic is exposed in one place, which aids debugging and extension but can be overwhelming to newcomers.

Quick start

Just visit CodeFlow and paste any GitHub URL.

Option 2: Self-Host

verdict

CodeFlow is a neat tool for developers and architects who want a quick, visual understanding of code structure and risk areas without setting up heavy analysis pipelines. Its zero-install, client-only design means you can use it anywhere with a modern browser and get immediate insights.

That said, the browser-based approach has limitations: it depends heavily on GitHub API availability and rate limits, and static analysis can only go so far without runtime context. Large repositories might challenge the visualization’s clarity and performance.

If you want a lightweight, privacy-conscious way to explore GitHub repos or your local codebases interactively, CodeFlow is worth trying. But for deep, runtime-aware code analysis, this isn’t the tool. It’s a clever, pragmatic solution that balances accessibility and feature richness with the inherent constraints of client-side processing.


→ GitHub Repo: braedonsaunders/codeflow ⭐ 2,826 · HTML