Version 2026.3.12
An autonomous AI agent with permanent memory & self-learning capabilities
😢 Traditional AI Assistants
"What do I like?"
"I don't know, you just told me."
Every conversation starts from scratch. They forget everything.
🤩 Resonix Remembers
"What do I like?"
"Based on our conversations, you prefer coffee over tea, usually take your coffee with oat milk, and you're a night owl who does your best work between 10PM-2AM."
Resonix actually remembers things. It learns from every conversation and builds on past knowledge.
✨ Key Features
🧠 Memory That Never Forgets
flowchart TB
USER["User"] -->|"message"| BOT["👾 Resonix"]
BOT --> LAYERA["Layer A"]
BOT --> LAYERB["Layer B"]
LAYERA --> PREF["Prefs"]
LAYERA --> FACTS["Facts"]
LAYERA --> PAT["Patterns"]
LAYERB --> ID["identity"]
LAYERB --> KNOW["knowledge"]
LAYERB --> LOGS["logs"]
style USER fill:#fff,stroke:#333,stroke-width:2,color:#000
style BOT fill:#7c3aed,stroke:#333,stroke-width:2,color:#fff
style LAYERA fill:#f3e8ff,stroke:#9333ea,stroke-width:2,color:#000
style LAYERB fill:#fdf2f8,stroke:#db2777,stroke-width:2,color:#000
style PREF fill:#fff,stroke:#333,stroke-width:1,color:#000
style FACTS fill:#fff,stroke:#333,stroke-width:1,color:#000
style PAT fill:#fff,stroke:#333,stroke-width:1,color:#000
style ID fill:#fff,stroke:#333,stroke-width:1,color:#000
style KNOW fill:#fff,stroke:#333,stroke-width:1,color:#000
style LOGS fill:#fff,stroke:#333,stroke-width:1,color:#000
Two-layer permanent memory. Layer A stores machine-readable profile, Layer B creates human-readable files at ~/Desktop/resonix-M/. Every conversation builds on the last.
🔄 Gets Smarter Over Time
flowchart LR
TASK["📋 Task"] --> RESULT["Result"]
RESULT --> RETRO["Review"]
RETRO --> LEARN["Learn"]
LEARN -.-> TASK
style TASK fill:#fff,stroke:#db2777,stroke-width:2,color:#000
style RESULT fill:#fff,stroke:#9333ea,stroke-width:2,color:#000
style RETRO fill:#fff,stroke:#7c3aed,stroke-width:2,color:#000
style LEARN fill:#fff,stroke:#6366f1,stroke-width:2,color:#000
Execute task → Analyze result → Create retrospective → Store learning → Improve next task. Remembers what worked and what didn't.
🌐 Browser That Just Works
flowchart LR
OLD["❌ Old"] -->|breaks| NEW["✅ Resonix"]
EXT["Ext"] -->|unreliable| PW["Playwright"]
SEL["Selector"] -->|unstable| SMART["Smart"]
SS["Manual"] -->|slow| AUTO["👾 Auto"]
style OLD fill:#fee2e2,stroke:#ef4444,stroke-width:2,color:#000
style NEW fill:#dcfce7,stroke:#22c55e,stroke-width:2,color:#000
style EXT fill:#fff,stroke:#333,stroke-width:1,color:#000
style PW fill:#fff,stroke:#22c55e,stroke-width:2,color:#000
style SEL fill:#fff,stroke:#333,stroke-width:1,color:#000
style SMART fill:#fff,stroke:#22c55e,stroke-width:2,color:#000
style SS fill:#fff,stroke:#333,stroke-width:1,color:#000
style AUTO fill:#fff,stroke:#22c55e,stroke-width:2,color:#000
Playwright-powered instead of fragile Chrome extensions. Smart element detection survives UI updates. Automatic screenshots & consistent results.
📊 Know Your Automation Status
pie title Task Results
"✅ Success 82%" : 82
"❌ Errors 18%" : 18
Success/Error rate tracking, P95 response time metrics, and risk level assessment. Always know how your automation is performing.
🏗️ System Architecture
flowchart TB
subgraph Users[Channels]
T[Telegram]
D[Discord]
S[Slack]
W[WhatsApp]
end
subgraph Gateway[Gateway]
R[Routing]
A[Auth]
C[Channels]
end
subgraph Agent[👾 Agent Runtime]
M[Memory]
B[Browser]
T2[Tools]
G[Growth]
I[Identity]
end
subgraph Memory[💾 Memory]
MA[Layer A]
MB[Layer B]
end
Users --> Gateway
Gateway --> Agent
Agent --> Memory
M --> MA
M --> MB
style Users fill:#fff,stroke:#7c3aed,stroke-width:2,color:#000
style Gateway fill:#fff,stroke:#7c3aed,stroke-width:2,color:#000
style Agent fill:#fff,stroke:#db2777,stroke-width:2,color:#000
style Memory fill:#fff,stroke:#9333ea,stroke-width:2,color:#000
style T fill:#fff,stroke:#7c3aed,stroke-width:1,color:#000
style D fill:#fff,stroke:#7c3aed,stroke-width:1,color:#000
style S fill:#fff,stroke:#7c3aed,stroke-width:1,color:#000
style W fill:#fff,stroke:#7c3aed,stroke-width:1,color:#000
style R fill:#fff,stroke:#333,stroke-width:1,color:#000
style A fill:#fff,stroke:#333,stroke-width:1,color:#000
style C fill:#fff,stroke:#333,stroke-width:1,color:#000
style M fill:#fff,stroke:#db2777,stroke-width:2,color:#000
style B fill:#fff,stroke:#9333ea,stroke-width:2,color:#000
style T2 fill:#fff,stroke:#333,stroke-width:1,color:#000
style G fill:#fff,stroke:#db2777,stroke-width:1,color:#000
style I fill:#fff,stroke:#db2777,stroke-width:1,color:#000
style MA fill:#fff,stroke:#9333ea,stroke-width:1,color:#000
style MB fill:#fff,stroke:#db2777,stroke-width:1,color:#000
🚀 Quick Start
One-Line Install
| Platform | Command |
|---|---|
| macOS / Linux | curl -fsSL https://raw.githubusercontent.com/mangiapanejohn-dev/Resonix-AG/main/install.sh | bash |
| Windows | iwr -useb https://raw.githubusercontent.com/mangiapanejohn-dev/Resonix-AG/main/install.ps1 | iex |
| npm | npm install -g github:mangiapanejohn-dev/Resonix-AG |
| Termux | curl -fsSL https://raw.githubusercontent.com/mangiapanejohn-dev/Resonix-AG/main/install-termux.sh | bash |
Get Started
resonix -v # Verify installation resonix onboard # First-time setup resonix gateway # Start the gateway
📱 Supported Channels
📂 What Resonix Remembers
~/Desktop/resonix-M/
│
├── identity/ # "Mark is a funny boy developer from Hong Kong"
├── knowledge/ # "Coffee with oat milk, not tea"
├── autonomy/ # "Learns faster when given examples"
├── retrospectives/ # "Don't use regex for HTML parsing"
└── logs/ # Conversation history
🎯 Resonix vs Traditional Assistants
| Feature | ChatGPT / Claude / Gemini | Resonix |
|---|---|---|
| Memory | Context window only | Permanent + Desktop mirror |
| Learning | Starts fresh each chat | Remembers & learns |
| Identity | Generic | Knows who you are |
| Browser | Extensions / API | Built-in Playwright |
| Persistence | Session-based | Long-term storage |
🛠️ Commands
| Command | Description |
|---|---|
resonix onboard |
First-time setup |
resonix gateway |
Start the gateway |
resonix gateway status |
Check status |
resonix cron board |
View scheduled tasks |
resonix memory profile |
View learned profile |
🤝 Contributing
# Clone the repo git clone https://github.com/mangiapanejohn-dev/Resonix-AG.git cd Resonix-AG # Install dependencies pnpm install # Build pnpm build # Test pnpm test
📚 Documentation
- Docs: docs.resonix.ai
- Discord: Join our community
- X: Follow us
⭐ Show Your Support
If Resonix helped you, we'd love a ⭐!
It motivates me to keep building and improving. 💪
📄 License
MIT License - Built by MarkEllington
Made with ❤️ by a Personl developer
