Reflyx

Documentation

Reflyx Documentation

Dual‑mode AI (Local + Cloud), semantic search, intelligent generation, and contextual explanations — all inside VS Code.

Overview

A production‑ready Reflyx exceeding the capabilities of Augment Code, Cursor, and Windsurf — with local privacy, cloud performance, semantic search, intelligent code generation, and contextual explanations.

At a glance

  • Privacy‑first, offline‑capable
  • Semantic search + fast inference
  • Local, Cloud, or Hybrid routing

Key Features

Dual‑Mode Processing

Seamlessly switch between Local (Ollama) and Cloud (GPT‑4o, Claude 3.5, Gemini, Groq).

Secure API Keys

Keys stored via VS Code SecretStorage for safe usage.

Ultra‑Fast Inference

Groq integration delivering 500+ tokens/sec for snappy responses.

Smart Routing

Intelligently select and fallback between providers for reliability.

Enhanced UI

Context‑aware chat, inline suggestions, and streaming responses.

Semantic Search

Index and search your workspace for relevant code instantly.

Supported Providers

Local (Free & Private)
  • Ollama: CodeLlama 7B/13B/34B, DeepSeek‑Coder 6.7B, Qwen2.5‑Coder 7B (16K‑32K context)
Cloud (Optional APIs)
  • OpenAI GPT‑4o (128K)
  • Anthropic Claude 3.5 Sonnet (200K)
  • Google Gemini 1.5 Pro (2M)
  • Groq LLaMA 3.1 70B (ultra‑fast)

Quick Start

  1. 1

    Automated Setup

    git clone <repo-url>
    cd ai-coding-assistant
    python setup.py
  2. 2

    Install VS Code Extension

    cd extension
    npm install
    npm run compile
    npm run package
  3. 3

    Configure Providers

    Choose Local, Cloud, or Hybrid inside VS Code settings.

Usage

  • Ctrl+Shift+P → Reflyx: Ask Codebase
  • Ctrl+Shift+P → Explain Selection
  • Ctrl+Shift+P → Generate Code
  • Ctrl+Shift+C → Toggle Chat

Configuration

{
  "aiCodingAssistant.modelProvider": "ollama",
  "aiCodingAssistant.embeddingModel": "all-MiniLM-L6-v2",
  "aiCodingAssistant.maxChunkSize": 500,
  "aiCodingAssistant.retrievalCount": 10
}

Troubleshooting

  • Ollama not found: ensure it's running (curl http://localhost:11434/api/tags).
  • Qdrant error: check status (curl http://localhost:6333/health).
  • Extension not loading: restart VS Code and check Developer Console.

Contributing

PRs welcome — features, fixes, and documentation improvements.

License

MIT — free for personal and commercial use.