🧠 AegisMind - Multi-Model AI Coding AssistantThe ONLY coding assistant that validates answers across GPT-4, Claude, and Mistral simultaneously. Stop trusting a single AI. AegisMind consults multiple models, detects bias, validates responses, and shows you which model contributed what. ✨ Features
🚀 Quick Start1. Install the ExtensionSearch for "AegisMind" in VS Code/Cursor Extensions marketplace. 2. Get Your Free AccountVisit aegismind.app and sign up (free tier: 10 queries/day). 3. Set Your EmailPress
4. Start Using It!Select any code → Right-click → Choose an AegisMind command! 💻 Code Commands📖 Explain Code (
|
| Command | Windows/Linux | Mac |
|---|---|---|
| Ask Question | Ctrl+Shift+A |
Cmd+Shift+A |
| Explain Code (selection) | Ctrl+Shift+E |
Cmd+Shift+E |
| Fix Code (selection) | Ctrl+Shift+F |
Cmd+Shift+F |
| Generate Tests (selection) | Ctrl+Shift+T |
Cmd+Shift+T |
Other commands available via right-click context menu
🎯 Context Menu
Select any code → Right-click to see:
┌─────────────────────────────────────┐
│ Cut │
│ Copy │
│ Paste │
├─────────────────────────────────────┤
│ 💡 AegisMind: Explain Code │
│ 🔧 AegisMind: Fix Code │
│ 🧪 AegisMind: Generate Tests │
│ ⚡ AegisMind: Optimize Code │
│ 📝 AegisMind: Add Comments │
│ 🔄 AegisMind: Refactor Code │
└─────────────────────────────────────┘
🎨 Transparent Multi-Model Provenance
Every response shows you exactly which models were consulted!
Provenance Banner (Top of Every Response)
> 🎨 Models Used for This Response:
>
> 🟢 GPT-4o
> 🔵 Claude-3.5
> 🟣 Mistral Large
This appears before the answer, so you always know which models contributed—even when synthesis is used.
Detailed Breakdown (Bottom of Response)
### 🎯 Original Models Consulted
- 🟢 GPT-4o (OpenAI)
- 🔵 Claude-3.5 (Anthropic)
- 🟣 Mistral Large (Mistral AI)
Total Models: 3 AI models consulted
Synthesis Method: Rule-based synthesis (consensus algorithm)
Validation Rounds: 2
Quality Metrics
- Coherence: 92% - Internal consistency
- Model Agreement: 87% - Consensus across models
Bias Detection
⚠️ Bias Correction: Bias detected and corrected during synthesis
The synthesis doesn't hide the source models—we show them prominently at the top!
🌍 19 Languages Supported
🇬🇧 English • 🇪🇸 Spanish • 🇫🇷 French • 🇩🇪 German • 🇵🇹 Portuguese
🇮🇹 Italian • 🇳🇱 Dutch • 🇫🇮 Finnish • 🇵🇱 Polish • 🇹🇷 Turkish
🇨🇳 Chinese • 🇯🇵 Japanese • 🇰🇷 Korean • 🇸🇦 Arabic • 🇷🇺 Russian
🇮🇳 Hindi • 🇬🇷 Greek • 🇮🇱 Hebrew • 🇹🇭 Thai
Auto-detection: Just ask questions or add code comments in your language!
⚙️ Settings
Ctrl+, → Search for "AegisMind":
{
"aegismind.email": "your@email.com",
"aegismind.defaultMode": "balanced",
"aegismind.showProvenance": true,
"aegismind.enableBiasDetection": true
}
Modes:
balanced(default) - Balance consensus and diversitycreative- More creative, exploratory responsesvalidated- Focus on verified, consensus answers
💰 Pricing
Free Tier Available
- ✅ Access to all AI models (GPT-4o, Claude, Mistral)
- ✅ Bias detection & validation included
- ✅ 19 languages supported
- ✅ All code commands available
Upgrade for More
Visit aegismind.app/billing for current pricing and plans.
Plans include:
- Higher query limits
- Faster response times
- Priority support
- API access (Enterprise)
- Custom model configurations
🔐 Privacy & Security
- ✅ Code used only for immediate AI inference
- ✅ Not stored permanently
- ✅ SOC 2 compliant infrastructure
- ✅ End-to-end encryption
🐛 Troubleshooting
"Please set your email first"
Run: Ctrl+Shift+P → AegisMind: Set Email
"403 Forbidden" or "Email not verified"
Check your email inbox for the verification link from AegisMind.
"Usage limit exceeded"
You've hit your free tier limit. Upgrade at aegismind.app/billing.
Commands not appearing?
Make sure you've selected code first - code commands only appear when text is selected.
Extension not activating?
Reload window: Ctrl+Shift+P → Developer: Reload Window
🔥 Corporate Firewall Issues
Problem: Extension times out or fails to connect from corporate networks.
Solution: The extension now routes through aegismind.app (instead of direct API calls) for firewall compatibility.
Default configuration (v1.3.0+):
{
"aegismind.apiEndpoint": "https://aegismind.app/api"
}
If you still have issues:
- Ask IT to whitelist:
aegismind.app(single domain) - Or manually set endpoint in settings:
- Open:
Ctrl+,→ Search "AegisMind" - Set API Endpoint:
https://aegismind.app/api
- Open:
Why this works:
- ✅ Corporate firewalls usually whitelist the web UI domain
- ✅ Requests are proxied through the UI (encoded/transformed)
- ✅ Bypasses Render WAF rules
- ✅ Same domain as web interface = no additional whitelist needed
🚀 Why AegisMind?
vs. GitHub Copilot
- ✅ Multi-model validation (vs. single OpenAI model)
- ✅ Bias detection (vs. black box)
- ✅ Provenance tracking (see which model contributed)
- ✅ 19 languages optimized (vs. English-first)
vs. Cursor Composer
- ✅ Free tier available (vs. $20/mo minimum)
- ✅ Transparent AI (vs. proprietary)
- ✅ Multi-model validation (vs. single model)
vs. ChatGPT
- ✅ Multi-model consensus (vs. single model)
- ✅ Code-optimized prompts (vs. general purpose)
- ✅ Bias detection built-in
- ✅ Integrated in your editor (vs. browser tab)
📖 Documentation
- Usage Guide: USAGE_GUIDE.md
- Publishing Guide: PUBLISH_TO_MARKETPLACE.md
- Changelog: CHANGELOG.md
📞 Support
- 📧 Email: support@aegismind.app
- 🌐 Website: aegismind.app
- 💬 GitHub Issues: github.com/tradingjohn/aegismind/issues
🎉 Examples to Try Right Now
1. Explain Complex Code
# Select this line and press Ctrl+Shift+E
list(map(lambda x: x**2, filter(lambda x: x % 2 == 0, range(10))))
2. Fix Buggy Code
// Select and press Ctrl+Shift+F
async function fetchData() {
return fetch('/api').then(data => data.json()); // Missing error handling!
}
3. Generate Tests
// Select and press Ctrl+Shift+T
func Add(a, b int) int {
return a + b
}
4. Optimize Performance
# Right-click → "Optimize Code"
result = []
for i in range(1000000):
if i % 2 == 0:
result.append(i * 2)
📄 License
MIT License - see LICENSE file
🎉 Built by Developers, for Developers
We believe AI should be transparent, validated, and bias-free.
Happy coding! 🚀