Complete Codebase Context
For AI-Assisted Development

$ rohkun run
Analyzing codebase...
Mapping architecture...
✓ Complete. 30 seconds.
Found 47 endpoints, 32 API calls
Architecture map copied to clipboard
✓ Paste into AI. Full context provided.
→ Run after edits to track changes
✓ Maintain clarity at scale

AI assistants lose context as codebases grow

Edits drift from architecture over time

Small mistakes compound into structural problems

Rohkun maintains continuity and scope

Start Scanning Free tier • No credit card required

Framework-Aware Analysis

AST-powered scanning across all major languages and frameworks

Python
JavaScript
TypeScript
Java
Go
Ruby
PHP
C#
Rust
Kotlin
Python
JavaScript
TypeScript
Java
Go
Ruby
PHP
C#
Rust
Kotlin

The Challenge of Scale

🔍

Limited Context Window

AI assistants can't see your entire architecture. They work with fragments, missing critical connections.

🧠

No Change Memory

Every edit happens in isolation. AI doesn't track what changed, why it changed, or where it's heading.

📉

Intent Drift

Small reasoning errors accumulate. Your architecture gradually diverges from its original purpose.

Architecture Visibility for AI

🔗
Endpoints
📡
API Calls
📸
Snapshots
📁
Dependencies

Comprehensive architecture mapping keeps AI aligned with your system's structure and intent.

Track structural changes over time to maintain continuity and prevent architectural drift.

Core Capabilities

🗺️

Complete Architecture Mapping

Single scan captures your entire codebase structure. Every endpoint and API call mapped with file locations and line numbers. High accuracy for standard code patterns. AI sees the full picture from the start.

📊

Structural Change Tracking

Run before and after AI edits to compute graph diffs. Track exactly what changed in your architecture over time. Maintain continuity across iterations.

📈

Drift Detection

Health score quantifies structural divergence from baseline. Catch architectural drift early before small errors compound. Keep AI aligned with your intent.

Efficient Token Usage

Provide complete context upfront instead of iterative explanations. Reduce failed attempts and rework cycles. Better results with fewer API calls.

🎯

Scope Awareness

AI understands the full extent of your system. Makes informed decisions considering dependencies and connections. Edits that respect your architecture.

How It Works

1

Install

pip install rohkun

Install the CLI tool

2

Run

rohkun run

Run analysis in your project directory

3

Report generated

Ctrl+V

Analysis results auto added to clipboard

With and Without Context

The difference between working blind and working with complete visibility

Rohkun provides the architectural awareness AI assistants need to maintain code quality at scale

Without Rohkun
Limited context, incremental drift
With Rohkun
Complete visibility, tracked changes
Fragmented View
AI works with limited context window. Can't see how your entire system connects. Makes decisions without full architectural awareness.
Complete Architecture Map
Single scan captures entire codebase structure. AI sees all endpoints, API calls, and dependencies at once. Makes informed decisions with full context.
No Change Tracking
Each edit happens in isolation. No record of what changed or why. Pattern of modifications lost over time.
Structural Change History
Run before and after edits to generate graph diffs. Track architectural evolution. Maintain continuity across development iterations.
Undetected Drift
Small reasoning errors accumulate silently. Architecture gradually diverges from original intent without visibility.
Early Detection
Health score quantifies structural divergence. Catch drift before it compounds. Keep architecture aligned with purpose.
Iterative Corrections
Multiple attempts to get changes right. Rework cycles waste time and tokens. Quality degrades with scale.
First-Pass Accuracy
AI makes better decisions with complete context. Fewer iterations needed. Quality maintained as codebase scales.
Manual Context Building
Repeatedly explain your architecture. Context gets lost between sessions. Productivity slows as complexity grows.
Automated Context Sharing
Single paste provides complete architectural context. Consistent understanding across sessions. Development velocity maintained.

Transparent Pricing

Pro Plan

Individual developers and small teams

$12.42/month

Billed annually

Total: $149/year

Monthly Credits
100 credits

Approximately 100 scans per month

Overage Rate
$0.10 / credit

Pay only for what you use beyond cap

Frequently Asked Questions

Rohkun performs AST-powered static analysis to map your codebase architecture. It identifies all endpoints and API calls with precise file locations and line numbers. The analysis covers backend routes, frontend API calls, WebSocket connections, GraphQL schemas, and library integrations across all major languages and frameworks. Each scan creates a snapshot that tracks structural changes over time, enabling drift detection and continuity tracking.

While Rohkun is optimized for AI-assisted development, it's valuable for any developer who needs clear visibility into their codebase architecture. The maps, change tracking, and drift detection work whether you code manually or with AI assistance. Teams use it for onboarding, documentation, and architectural reviews.

Most codebases complete analysis in under 60 seconds. Large projects with hundreds of files typically finish in 2-3 minutes. The CLI runs locally, creates a snapshot, then uploads a compressed package for server-side processing. Run again after edits to generate comparison reports showing exactly what changed.

Rohkun supports Node.js (Express, Fastify, Nest), Python (FastAPI, Flask, Django), PHP (Laravel), Java (Spring Boot), Go (Gin, Echo), Ruby (Rails, Sinatra), C# (ASP.NET), Rust (Actix, Rocket), and Kotlin (Ktor). Frontend frameworks include React, Vue, Angular, Svelte, and Next.js. Also covers GraphQL, REST, WebSockets, and gRPC.

Your codebase is uploaded as a compressed archive to our secure servers for analysis. We generate the architecture map and store it for your reference. Source code can be automatically deleted after processing if you prefer. All data is encrypted in transit and at rest. You control retention policies through your dashboard.

Rohkun continues working seamlessly with pay-as-you-go overage charges. Pro Plan pays $0.10 per additional credit, Premium Plan pays $0.08 per credit. No service interruption, no artificial limits. You only pay for actual usage beyond your monthly allocation.

Cancel anytime from your account dashboard. Plans are billed upfront but cancellation prevents automatic renewal. You retain full access through the end of your current billing period. No penalties, no questions asked.

Formula: (Endpoints + API Calls) × 100 tokens per item. Each architectural element requires roughly 100 tokens to describe to an AI assistant (endpoint path, HTTP method, file location, parameters). Instead of explaining manually, you paste the complete map once. A project with 50 endpoints and 40 API calls saves approximately 9,000 tokens per context (about $0.27 at GPT-4 pricing). Actual savings multiply when you factor in reduced rework from better first-pass accuracy.

AST-powered analysis provides high accuracy for standard code patterns. Endpoint detection is highly reliable across supported frameworks. API call detection works well for typical usage patterns. Complex or unconventional code may require manual verification, which is why we provide exact file locations and line numbers for every detected element. We focus on deterministic analysis rather than heuristic guessing. Connection matching between frontend and backend is pattern-based and may require manual verification for complex routing.

Simple: (1) Sign up for a free account, (2) Create an API key from your dashboard, (3) Install the CLI with pip install rohkun, (4) Run rohkun analyze --api-key YOUR_KEY in your project directory. You'll get a complete analysis report in seconds.

API keys authenticate your CLI tool with your account. Create multiple keys for different machines (laptop, desktop, CI/CD) or team members. Each key tracks its own usage statistics while sharing your account's credit pool. You can name them (e.g., "Alice's Laptop", "Production CI") and revoke individual keys without affecting others.

Yes. One account can have multiple API keys for different team members. All usage counts against the account's credit pool, making billing simple. Each key shows its own usage stats so you can track who's using what. Perfect for small teams that want centralized billing without complex permission systems.

Results appear immediately in your terminal after analysis completes. You can also view all your projects in the web dashboard, where you can filter, search, and bookmark important analyses. Each project shows detected endpoints, API calls, frameworks, and file statistics.

Yes. Revoke or delete any API key independently from your dashboard. The key immediately stops working while all other keys remain active. Useful when team members leave, devices are lost, or for routine security key rotation. Historical usage data and statistics remain visible for auditing purposes.