Understand Any Codebase In Minutes
Self-hosted AI documentation assistant that transforms how developers understand code. RAG-powered chat, intelligent code audit, dependency visualization, and file exploration - all running privately on your machine.
🚀 Perfect for New Team Members
New developer at a company? Index the codebase and ask questions like "How does authentication work?" or "Where is the payment logic?" - onboard in hours, not weeks.
middleware/auth.ts validates JWT tokens and attaches user data to the request context...Powerful Features for Modern Development Teams
Everything you need to understand, audit, and document any codebase - with complete privacy
100% Self-Hosted & Private
Your code NEVER leaves your machine. Everything runs locally via Docker - no cloud, no data sharing, complete privacy.
5 AI Models Supported
GPT-4o, GPT-4o Mini, GPT-4 Turbo, Claude 3.5 Sonnet, Claude 3 Haiku. Switch models anytime from the UI.
Smart RAG with LanceDB
Semantic vector search finds the most relevant code chunks. Ask anything - get answers with exact file:line references.
Persistent Chat Sessions
Create multiple chat sessions per project. History is saved - continue where you left off, never lose context.
Private GitHub Repos
Import private repositories with Personal Access Tokens. Enterprise-ready for your company's private codebases.
AI Code Audit (3 Levels)
🔴 Security (SQL injection, XSS, hardcoded keys) 🟡 Performance (N+1, memory leaks) 🟢 Best Practices with fix suggestions.
Interactive Dependency Graph
Visual map of all imports and file relationships. Supports 40+ languages. Understand architecture at a glance.
VS Code-like File Explorer
Browse project files with syntax highlighting. Click any file to view code with proper highlighting for all languages.
Global Code Search
Search across all files instantly. Find functions, classes, variables - navigate large codebases effortlessly.
File References in Answers
Every AI response includes clickable file paths and line numbers. Jump directly to the relevant code.
Real-time Streaming
See AI responses as they're generated - no waiting. Smooth, fast streaming like ChatGPT.
Fast Onboarding for Teams
New developer? Index the project, ask questions, understand the codebase in hours instead of weeks.
PDF Report Export
Export Code Audit and Project Health analyses as professional PDF reports — perfect for client handoffs or team reviews.
Live Code Analyzer
Paste any code and get instant analysis with highlighted lines, severity scores, and side-by-side fix suggestions.
PR Auto-Reviewer
Coming SoonGitHub & GitLab webhook integration that automatically reviews pull requests and posts AI code review comments.
Built for Real Developer Workflows
See how teams use DocuBrain to save hours every week
New Team Member Onboarding
Day 1 at a new company? Index the codebase and start asking questions. Understand authentication flow, database schema, API structure - all in hours, not weeks.
"How does the payment processing work?" → Get detailed explanation with file references
Code Review & Security Audit
Run AI-powered audits to find SQL injection, XSS vulnerabilities, hardcoded credentials, and performance issues before they hit production.
Automatically detects: eval(), innerHTML, hardcoded API keys, N+1 queries
Legacy Code Documentation
Inherited a codebase with zero documentation? DocuBrain becomes your documentation. Ask what any function does and get clear explanations.
"What does processUserData() do?" → Full breakdown with dependencies
Architecture Understanding
See how files connect with the dependency graph. Understand which components depend on what before making changes.
Visual graph showing: App → Components → Services → Database
See DocuBrain In Action
Six powerful views to understand, audit, explore, analyze, and monitor your codebase
The RAGService in backend/app/services/rag.py handles retrieval-augmented generation for code Q&A.
Here's how it works:
- Takes your question and the project ID
- Uses LanceDB to find semantically similar code chunks
- Builds a context-aware prompt with relevant code snippets
- Streams the LLM response back to you
Quick Setup & Easy Installation
After purchase, you'll receive the complete application. Just configure and run with Docker.
1. Download & Extract
After purchase, download the ZIP file and extract it to your preferred location.
2. Configure API Keys
Add your OpenAI or Anthropic API keys to the .env file. Both providers are supported.
3. Start with Docker
Run docker-compose up -d and open localhost:3000. That's it!
Two Ways to Add Your Projects
Import codebases from your local machine or clone directly from GitHub - public or private repos
Local Folder Import
For projects already on your machine
Click "Add Project" in sidebar
Opens the project import dialog
Select "Local Folder" tab
Choose from your file system
Browse and select your project folder
Any folder with source code
DocuBrain indexes automatically
Files are processed and ready for chat
💡 Perfect for: Work projects, personal repos, any local codebase you're developing
GitHub Repository Clone
Public & private repos supported
Click "Add Project" in sidebar
Opens the project import dialog
Select "GitHub" tab
Enter repository URL
Paste repo URL (add token for private repos)
e.g., github.com/company/private-app
DocuBrain clones and indexes
Repo downloaded and processed locally
🔒 Private repos: Add your GitHub Personal Access Token in Settings for private repository access
The "Add Project" dialog makes it easy:
Built with Modern Tech Stack
Enterprise-grade technologies for reliability and performance
Simple, One-Time Pricing
No subscriptions. Pay once, own forever. Your code, your control.
Standard License
For individuals & small teams
- Complete working application
- Docker deployment ready
- Unlimited repositories
- All 5 AI models (GPT-4o, Claude 3.5, etc.)
- RAG Chat + Code Audit + Dependencies + File Explorer
- Private GitHub repo support
- Persistent chat sessions
- Lifetime updates & bug fixes
- Email support
Source Code License
Full customization power
- Full source code (Next.js + FastAPI + Python)
- Modify UI, add features, change logic
- Add your own AI models
- Integrate with your existing systems
- White-label for clients
- Commercial usage rights
- Priority support
- Private GitHub repository access
Frequently Asked Questions
How does DocuBrain keep my code private?
DocuBrain runs 100% on your local machine using Docker. Your code is indexed locally in LanceDB, and API calls to OpenAI/Anthropic only contain relevant code snippets - never your entire codebase.
Which AI models are supported?
GPT-4o, GPT-4o Mini, GPT-4 Turbo (OpenAI), Claude 3.5 Sonnet, Claude 3 Haiku (Anthropic), and Gemini 2.0 Flash, Gemini 1.5 Pro (Google). You can switch between models anytime from the UI.
Do I need my own API keys?
Yes, you'll need your own OpenAI or Anthropic API keys. You pay for AI usage directly to the providers at their standard rates.
Can I use this with private repositories?
Yes! DocuBrain supports private GitHub repositories. Just provide your Personal Access Token when adding a project.
What are the system requirements?
Docker and Docker Compose. Minimum 4GB RAM, 8GB+ recommended. Works on Windows, macOS, and Linux.
What's the difference between Standard and Source Code licenses?
Standard gives you the full working application. Source Code includes everything plus full access to modify, customize, and extend the codebase.
Ready to Understand Your Codebase?
Start chatting with your code today. Self-hosted, private, and powerful.