🖥️ AI Detector — Can My PC Run This AI Model?

Instantly check which local AI large language models your computer can run

🚀 Live Demo · 📖 Usage Guide · 🤝 Contributing · 💬 Issues · 中文文档

---

✨ Overview

AI Detector is a free, browser-based hardware detection tool. It automatically reads your device's RAM, CPU, and GPU information, intelligently matches compatibility with 21+ mainstream open-source AI models, and estimates local inference speed — helping you find the best AI model for your machine in under 5 seconds.

No installation. No login. No data upload. 100% runs locally in your browser.

---

🎯 Key Features

🔍 Automatic Hardware Detection

• System RAM — reads via navigator.deviceMemory
• CPU Cores — reads logical core count via navigator.hardwareConcurrency
• GPU Model — detects via WebGL WEBGL_debug_renderer_info extension
• AI Score — calculates a 0–100 composite score to rank your device's AI capability level

🤖 Compatibility Analysis for 21+ LLMs

Covers today's most popular open-source model families:

Family	Representative Models	Parameter Range
🦙 Llama	TinyLlama, Llama 3.2, Llama 3.1	1.1B ~ 70B
🌐 Qwen	Qwen2.5 3B/7B/14B/32B/72B	3B ~ 72B
💎 Phi	Phi-3 Mini, Phi-3 Medium	3.8B ~ 14B
🌬️ Mistral	Mistral 7B	7B
🧠 DeepSeek	DeepSeek-R1, DeepSeek Coder	7B ~ 70B
👁️ Multimodal	LLaVA 7B, MiniCPM-V 8B	7B ~ 8B
💫 Gemma	Gemma 2 2B	2B

⚡ Three-Tier Compatibility Labels

• ✅ Runs Smoothly — sufficient RAM for stable, fast inference
• ⚠️ Barely Runs — just meets the minimum, expect slower speeds
• ❌ Not Supported — insufficient memory to load this model

📊 Speed Ranking (tokens/s Estimate)

After detection, a performance leaderboard is auto-generated, ranking all runnable models by estimated tokens per second.

🔧 Quantization Mode Toggle

• Q4 — lower memory footprint, recommended for most devices
• Q8 — higher precision, requires ~2× the RAM of Q4
• Switch in one click; all compatibility statuses update in real time

🎁 Personalized Model Recommendations

Based on your hardware profile, the tool automatically selects the most balanced, fastest, and most capable models for you.

📋 One-Click Report Copy

Generates a shareable text report with your hardware specs, score, and runnable model list.

---

🚀 Quick Start

Option 1 — Use the Online Version

Visit https://jitword.com/ai-detector. No setup required.

Option 2 — Run Locally

# 1. Clone the repo
git clone https://github.com/MrXujiang/ai-detector.git

# 2. Enter the directory
cd ai-detector

# 3. Start with any static server (e.g. npx)
npx serve .
# Or simply open ai-detector.html directly in your browser

The project consists of just 3 files (HTML / CSS / JS). Zero dependencies, zero build steps.

---

📖 Usage Guide

1. Open the page → click "Start Detection"
2. Wait ~3–5 seconds → RAM, CPU, and GPU are scanned automatically
3. View results → see your AI score and device tier
4. Browse recommendations → the "Recommended for You" section shows your top 3 models
5. Filter models → filter by compatibility status or category; search by keyword
6. Switch quantization → toggle Q4 / Q8 to see how memory requirements change
7. Copy report → one-click copy of your full detection result

---

💡 Why AI Detector?

Feature	AI Detector	Other Tools
Installation required	❌ None	✅ Usually
Login required	❌ None	✅ Usually
Data upload	❌ Never	⚠️ Sometimes
Model coverage	✅ 21+ mainstream models	⚠️ Limited
Speed estimation	✅ tokens/s quantified	⚠️ Rarely available
Upgrade suggestions	✅ Smart RAM advice	❌ No
Quantization toggle	✅ Q4 / Q8 live switch	❌ No
Open source & free	✅ MIT License	⚠️ Mostly paid

---

🛠️ Recommended Local Inference Tools

Once you know which models your device supports, use these tools to run them locally:

🖥️ Ollama — Best for Beginners

One command to download and run any model. Supports macOS / Linux / Windows. 300+ models available:

ollama run qwen2.5

🖱️ LM Studio — Zero-Code Friendly

Desktop GUI app with a built-in chat interface. No command line needed. Search and download models from HuggingFace directly.

⚙️ llama.cpp — Developer's Choice

Pure C++ implementation with the lowest memory footprint. Supports CPU/GPU hybrid inference. Powers Ollama under the hood:

./llama-cli -m model.gguf -p "Hello!"

---

📁 Project Structure

ai-detector/
├── ai-detector.html   # Main page (HTML structure + SEO metadata)
├── ai-detector.css    # Styles (animations, cards, nav, responsive layout)
├── ai-detector.js     # Core logic (hardware detection, model DB, rendering)
├── README.md          # English documentation (this file)
└── README_CN.md       # Chinese documentation

---

🔒 Privacy

This tool takes privacy seriously:

• All detection runs entirely in-browser using native JavaScript APIs
• No data is ever sent to any server
• No cookies, no user tracking
• RAM values are intentionally approximate due to browser privacy protections (e.g. 2, 4, 8, 16, 32 GB)

---

🤝 Contributing

Contributions are welcome! Feel free to open Issues and Pull Requests.

• 🐛 Found a bug? Open an Issue
• 💡 Have an idea? Start a Discussion
• 🆕 Want to add a model? Edit the MODELS array in ai-detector.js and submit a PR

# Fork and clone your copy
git clone https://github.com/your-username/ai-detector.git

# Create a feature branch
git checkout -b feature/add-new-model

# Commit your changes
git commit -m "feat: add support for xxx model"

# Push and open a PR
git push origin feature/add-new-model

---

🌟 Related Products

More AI tools built and maintained by the author:

Product	Description	Link
JitWord	Collaborative AI document editor · Real-time co-editing · Knowledge base	jitword.com
JitKnow	RAG-powered AI knowledge base · Private deployment · Smart Q&A	know.jitword.com
Pxcharts	AI-powered spreadsheets · Data visualization · Online collaboration	pxcharts.turntip.cn

---

📬 Contact

• GitHub: @MrXujiang
• WeChat Official Account: Scan the QR code in the bottom-right corner of the page for more AI tools and open-source resources
• Issues: github.com/MrXujiang/ai-detector/issues

If this project helps you, a ⭐ Star is the best encouragement!

---

Made with ❤️ by MrXujiang

Help every computer find the AI model that fits it best.