MAESTRO Documentation¶

Your Self-Hosted AI Research Assistant

Welcome to MAESTRO¶

What is MAESTRO?

MAESTRO is an AI-powered research platform you can host on your own hardware. It's designed to manage complex research tasks from start to finish in a collaborative research environment.

Transform your research workflow: Plan your research questions → Let AI agents investigate → Receive comprehensive reports with citations

Why Choose MAESTRO?¶

Advanced AI Integration
Leverage multiple LLMs including GPT-5, Claude, and local models for diverse research capabilities

Deep Document Analysis
Advanced RAG pipeline with dual embeddings for superior search accuracy

Iterative Research Process
Multi-agent system that refines findings through reflection and critique loops

Complete Data Privacy
Self-hosted solution ensures your sensitive research never leaves your infrastructure

Quick Links¶

Quick Start

Get MAESTRO up and running in minutes with Docker
Installation

Detailed installation instructions for various platforms
User Guide

Learn how to use all of MAESTRO's features
Example Reports

See real research outputs from various AI models

Core Features¶

Document Management

Upload and manage your research documents in a central library
- PDF, Word, and Markdown support
- Advanced RAG pipeline with BGE-M3 embeddings
- Semantic search across all documents
- Document groups for organized research
Intelligent Research

AI agents conduct thorough research based on your requirements
- Multi-agent system with 5 specialized agents
- Iterative research with reflection loops
- Automatic source citation and tracking
- Customizable research depth and focus
Writing Assistant

AI-powered writing support for reports and documentation
- Context-aware suggestions from your library
- Real-time collaborative editor
- LaTeX formula support
- Reference management and citations
Web Integration

Seamlessly search and fetch content from the internet
- Multiple search providers (Tavily, LinkUp, Jina)
- Automatic content extraction and parsing
- JavaScript-heavy site support
- Smart query expansion
Multi-User Support

Built for collaborative research environments
- Role-based access control
- Personal document libraries
- Shared research missions
- Individual settings and preferences
Self-Hosted Privacy

Complete control over your data and infrastructure
- Run on your own hardware
- No data leaves your servers
- Support for local LLMs
- Full audit trail and logging

System Requirements¶

Minimum Requirements¶

Docker & Docker Compose v2.0+
16GB RAM
30GB free disk space
Internet connection
API keys for at least one AI provider

Recommended Setup¶

32GB+ RAM for better performance
50GB+ SSD storage
NVIDIA GPU for 3-5x faster processing
Multiple AI provider keys
Local LLM capability (RTX 3090/4090)

Architecture Overview¶

MAESTRO is built with a modern, scalable architecture:

graph TB
    subgraph "Client"
        A[React Frontend]
    end

    subgraph "Gateway"
        B[Nginx Reverse Proxy]
    end

    subgraph "Backend Services"
        C[FastAPI Server]
        D[Multi-Agent System<br/>Controller • Planning • Research<br/>Reflection • Writing]
        E[RAG Pipeline<br/>BGE-M3 Embeddings]
    end

    subgraph "Data Layer"
        F[(PostgreSQL + pgvector<br/>Documents • Vectors • Chats)]
        G[File Storage<br/>PDFs • Markdown]
    end

    subgraph "External"
        H[AI Providers<br/>OpenAI • Anthropic • Local LLMs]
        I[Search Providers<br/>Tavily • LinkUp • Jina]
    end

    A --> B
    B --> C
    C --> D
    C --> E
    D --> E
    E --> F
    C --> F
    C --> G
    D --> H
    D --> I

    style F fill:#e1f5fe,stroke:#01579b,stroke-width:2px
    style D fill:#f3e5f5,stroke:#4a148c,stroke-width:2px

Documentation Sections¶

Getting Started¶

Quick Start - Get up and running in minutes
Installation - Platform-specific installation guides
Configuration - Configure AI providers and settings
First Login - Initial setup and user creation

Using MAESTRO¶

User Guide - Complete guide to all features
Research - Creating and managing research missions
Writing - Using the AI writing assistant
Documents - Managing your document library
Settings - Customizing your experience

Advanced Topics¶

Architecture - System design and components
Local LLM Deployment - Running models on your hardware
Example Reports - Sample outputs from various models
Troubleshooting - Solutions to common issues

Latest Updates¶

Version 0.1.10-alpha Released Oct 12, 2025

Azure OpenAI & Configuration Improvements
- Full Azure OpenAI support including GPT-5 models with automatic parameter handling
- Manual model entry toggle for providers without /models endpoint support
- Fixed 401 errors from external providers no longer logging users out
- Mission settings now persist correctly across server restarts
- Disabled autocomplete on API key fields to prevent browser autofill
Version 0.1.9-alpha Released Oct 3, 2025

Stability & Security Update
- Mission stability with proper checkpoint handling
- Security update replacing passlib with maintained libpass fork
- Bug fixes for Round/Pass counter and activity log persistence
- Password compatibility fixed for authentication
Version 0.1.8-alpha Released Sep 26, 2025

Mission Resilience & Document Intelligence Update
- Intelligent mission resume with complete checkpoint preservation
- arXiv paper fetcher for direct academic paper processing
- Writing phase resume support for interrupted missions
- Document reprocessing and re-embedding capabilities
Version 0.1.7-alpha Released Jan 25, 2025

Intelligent Document Management & Enhanced Research
- Auto-create document groups from web sources during research
- Seamless transition with report and document sources from research to writing workspace
- Research report versioning with easy version switching
- Clickable document sources in writing mode
- Restart and revise missions with improved outline handling
Version 0.1.6-alpha Released Jan 23, 2025

Model Support & Cost Tracking
- GPT-5 support with configurable thinking levels
- Improved cost tracking system
- DeepSeek compatibility improvements
- Enhanced error handling and retry logic
Version 0.1.5-alpha Released Sep 2, 2025

Major Performance & Stability Update
- Complete async backend migration (2-3x faster)
- Fixed mission resume and recovery capabilities
- Complete documentation overhaul
- Enhanced UI/UX with LaTeX support
- 50+ bug fixes and stability improvements

Join the Community¶

GitHub

View Repository

Source code, issues, and contributions
Discussions

Join Community

Ask questions and share experiences
Issues

Report Issues

Report bugs and request features

License¶

MAESTRO is open source software licensed under the GNU Affero General Public License v3.0, ensuring it remains free and open for everyone.

🚀 Ready to Transform Your Research?

🐳 Install with Docker ⚡ Quick Start Guide 📚 Read Documentation

Join hundreds of researchers using MAESTRO to accelerate their work