> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lostmindai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# LostMind AI - Project Analyser with RAG (Gemini) - Architecture

> architecture documentation for LostMind AI - Project Analyser with RAG (Gemini)

# Architecture Overview

## System Overview

Gemini Analyzer couples a Python FastAPI service with a Tkinter desktop client to deliver dual-mode project intelligence: cloud-powered analysis via Google Gemini 2.5 or fully local processing for sensitive codebases.

## Application Modes

* **Desktop GUI** — Tkinter interface that performs local scans and presents insights without leaving the workstation
* **REST API Service** — FastAPI deployment with Docker packaging for CI/CD and enterprise integrations
* **Shared Core Library** — Common analysis logic consumed by both the GUI and API layers

## Key Components

* **Analysis Engine**: Multi-language parsers, dependency mapping, and metrics collectors
* **AI Orchestrator**: Routes requests to Gemini 2.5 or privacy-first local models
* **Storage Layer**: Project metadata, cached insights, and optional SQL backend
* **Config & Validation**: Pydantic settings plus strict schema validation

## Integrations

* **Google Generative AI SDK** for hosted intelligence when API keys are provided
* **Docker** deployment for consistent production environments
* **CLI & Scripts** to bootstrap analysis tasks and health checks

## Security & Privacy

* Local-only mode ensures sensitive repositories never leave the machine
* Sanitised uploads for remote analysis with optional API key gating
* Audit-ready logging across both GUI and API deployments

*Summarised from the Gemini Analyzer project blueprint and README.*