⚡ SMARAN.AI Coder

🧠 Autonomous AI Pair Programmer • ⚡ Multi-LLM Orchestration • 📊 100% Genuine Host Telemetry

🌟 Two Ways to Use SMARAN.AI

🌐 Mode 1: Web Browser Studio (`http://localhost:3003`)

If you want the full standalone software engineering experience with Interactive Live HTML/CSS Preview, Document RAG, Live Web Research, and 100% Host Hardware Telemetry:

Run the Docker container on any machine:

docker run -d -p 3003:3003 --name smaran-ai-app shashwatmishra062/smaran-ai:latest

Open your web browser and visit: 👉 http://localhost:3003
Enjoy autonomous coding, document intelligence, and real hardware monitoring!

💻 Mode 2: In-Editor VS Code Extension

Use SMARAN.AI directly inside VS Code for real-time code generation, AST context analysis, and inline refactoring:

Open the SMARAN.AI Coder tab in the sidebar (🧠 icon).
If local server is running: Connects automatically to http://localhost:3003.
If using standalone cloud mode: Click Settings (⚙️) in the extension header and enter your free OpenRouter (sk-or-v1-...) or Groq (gsk_...) API key for direct cloud inference!

⚡ Multi-LLM Dynamic Routing Arsenal

🎯 19 Intelligent Routing Strategies → Priority, Weighted, Round-Robin, Cost-Optimized, Latency-First, Auto-Combo

🏷️ Provider	🤖 Model	🌟 Capability
🟣 DeepSeek	DeepSeek R1 Reasoning	◈ Ultra-deep mathematical reasoning & step-by-step logic
⚡ Groq Cloud	LLaMA 3.3 70B Versatile	◈ 500+ tokens/sec ultra-fast code generation
🔵 Google	Gemini 2.0 / 2.5 Flash	◈ High-speed multimodal code synthesis
🟠 OpenRouter	Community Free Routes	◈ Zero-cost community models with automatic failover
🟢 Nvidia NIM	Nemotron 70B Instruct	◈ Hardware-accelerated code and instruction following
🔴 Anthropic	Claude 3.5 Sonnet / Opus	◈ Enterprise-grade architecture & refactoring
💻 Local	Qwen 2.5 Coder / DeepSeek	◈ 100% offline private local inference via Ollama / vLLM

🗜️ Headroom Compression & Memory Vault

🚀 Headroom 60–90% Token Reduction: RTK filtering and AST context relay minimize token costs and optimize response latency.
🧠 Claude-Mem Long-Term Memory: Persistent episodic memory retains user preferences and project architecture across sessions.
🛡️ STRIX Security & Local Sandbox: Static AST code vulnerability scanning (SQLi CWE-89, IDOR, Secret Leakage) with zero telemetry leakage.

🎯 Quick Action Chips in VS Code

Chip	Action	What It Does
⚡	Explain	Deep structural and logical breakdown of selected code
🛠️	Refactor	Performance, readability, and clean architecture optimization
🧪	Unit Tests	Comprehensive test suite generation covering edge cases
🐞	Fix Bugs	Security vulnerability detection and bug repair

👨‍💻 About the Developer

Shashwat Mishra

AI Engineer & Robotics Engineer • Creator & Architect of SMARAN.AI

_{Built with ❤️ by Shashwat Mishra • SMARAN.AI v1.0.6 Production Extension}

SMARAN.AI - Autonomous Engineering Agent & Multi-LLM Copilot

Shashwat Mishra

⚡ SMARAN.AI Coder

🧠 Autonomous AI Pair Programmer • ⚡ Multi-LLM Orchestration • 📊 100% Genuine Host Telemetry

🌟 Two Ways to Use SMARAN.AI

🌐 Mode 1: Web Browser Studio (http://localhost:3003)

💻 Mode 2: In-Editor VS Code Extension

⚡ Multi-LLM Dynamic Routing Arsenal

🗜️ Headroom Compression & Memory Vault

🎯 Quick Action Chips in VS Code

👨‍💻 About the Developer

Shashwat Mishra

🌐 Mode 1: Web Browser Studio (`http://localhost:3003`)