Gaming Assistant AI

AI-powered in-game overlay — voice commands, screenshot analysis, build recommendations, RAG knowledge base.

Built with Tauri 2.0 + React + TypeScript. Runs as a lightweight always-on-top window alongside any game.

---

Features

Feature	Status	Powered By
AI Chat	✅	DeepSeek V3
Voice Commands (STT)	✅	Deepgram Nova-2
Voice Responses (TTS)	✅	OpenAI TTS
Screenshot Analysis	✅	Gemini Vision
Build Recommender	✅	RAG + DeepSeek
Game Detection	✅	Native (Windows)
RAG Knowledge Base	✅	Qdrant + OpenAI Embeddings
Cost Dashboard	✅	Built-in tracking

---

Screenshots

Add your own screenshots here

---

Quick Start

Prerequisites

• Node.js 18+
• Rust + Cargo
• Qdrant (optional, for RAG)

Install & Run

git clone https://github.com/your-username/gaming-assistant.git
cd gaming-assistant

npm install

cp .env.example .env
# Fill in your API keys in .env

npm run tauri dev

Build Installer

npm run tauri build
# Output: src-tauri/target/release/bundle/

---

API Keys

All keys go in .env (copy from .env.example).

Variable	Service	Free Tier	Get It
VITE_DEEPSEEK_API_KEY	LLM Chat	$0.14/1M tokens	platform.deepseek.com
VITE_GEMINI_API_KEY	Vision	Yes (limited)	aistudio.google.com
VITE_OPENAI_API_KEY	TTS + Embeddings	No	platform.openai.com
VITE_DEEPGRAM_API_KEY	Voice STT	$200 credit	console.deepgram.com
VITE_RAWG_API_KEY	Game DB	Free	rawg.io/apidocs

IGDB and Qdrant keys are optional — see .env.example for full list.

---

Usage

Chat

Type any question. AI pulls context from the RAG knowledge base, responds with cited answers.

Voice Commands

1. Go to Voice tab
2. Hold Space or click the mic
3. Speak — AI responds with voice

Screenshot Analysis

1. Alt+Tab back to your game, take mental note
2. Alt+Tab to the assistant
3. Click Capture Screenshot — Gemini Vision analyzes the screen

Build Recommender

Select game → class → playstyle → Find Builds for curated, RAG-backed recommendations.

RAG Knowledge Base (Optional)

# Start Qdrant locally
docker run -p 6333:6333 qdrant/qdrant

# Scrape wiki data (Python)
pip install requests beautifulsoup4 lxml
python scripts/scraper.py --game elden_ring --output data/

# Index into Qdrant
npx tsx scripts/index-knowledge.ts --input data/elden_ring.json

---

Architecture

gaming-assistant/
├── src-tauri/                  # Rust backend
│   ├── src/
│   │   ├── commands.rs         # Tauri IPC commands
│   │   ├── screenshot.rs       # Screen capture
│   │   └── game_detection.rs   # Active game detection
│   └── tauri.conf.json
├── src/
│   ├── components/             # React UI
│   │   ├── Chat.tsx
│   │   ├── VoiceInterface.tsx
│   │   ├── ScreenshotAnalyzer.tsx
│   │   ├── BuildRecommender.tsx
│   │   ├── CostDashboard.tsx
│   │   └── Sidebar.tsx
│   ├── services/               # API integrations
│   │   ├── deepseek-llm.ts
│   │   ├── gemini-vision.ts
│   │   ├── deepgram-stt.ts
│   │   ├── openai-tts.ts
│   │   ├── rag-pipeline.ts
│   │   └── vectordb.ts
│   ├── store/                  # Zustand state
│   └── hooks/
└── scripts/
    ├── scraper.py              # Wiki scraper
    └── index-knowledge.ts      # RAG indexer

---

Cost Estimates

Per interaction (approximate):

Action	Cost
Chat message	~$0.001
Voice command	~$0.008
Screenshot analysis	~$0.015

Monthly at 1,000 active users:

Service	Cost
DeepSeek V3 (chat)	~$36
Deepgram STT	~$130
OpenAI TTS	~$150
Gemini Vision	~$15
OpenAI Embeddings	~$5
Qdrant Cloud	~$10
Total	~$346 / mo

~$0.35 per user per month.

---

Development

npm test                  # Run unit tests
npm run test:coverage     # With coverage report
npm run test:e2e          # Playwright E2E
npm run lint              # ESLint

---

Contributing

See CONTRIBUTING.md.

---

Security

Found a vulnerability? See SECURITY.md.

---

License

Source code is available under the MIT License + Commons Clause.

• Free for personal use, self-hosting, learning, and contribution.
• Selling or commercially redistributing this software (or derivatives) requires written permission.

See LICENSE for full terms.

© 2025 BroHelp