Skip to main content
SHARC provides semantic code search capabilities through the Model Context Protocol (MCP). This guide will help you set up SHARC with your preferred AI assistant.

Prerequisites

Before you begin, ensure you have:
  • Node.js 18+ or Bun 1.0+ installed
  • A SHARC API key from your plan
  • An MCP-compatible AI assistant (Claude Code, Cursor, VS Code, etc.)

Quick Start

The fastest way to get started is with Claude Code:
claude mcp add sharc \
  -e SHARC_API_KEY=your_api_key \
  -- npx @sharc-code/mcp@latest
That’s it! Now you can index and search your codebases:
> Index this codebase for semantic search

● index_codebase (MCP)
  ⎿ Indexed 2,450 chunks from 342 files in 18.3s

> How does authentication work?

● search_code (MCP)
  ⎿ Found 3 results with scores 0.94-0.98

What’s Next?

How It Works

  1. Index your codebase - SHARC analyzes your code, splits it into semantic chunks, and creates embeddings
  2. Ask questions - Use natural language to search for code functionality
  3. Get precise results - Receive ranked code snippets with relevance scores
  4. Stay updated - File watching keeps your index current as you code

Performance

SHARC supports two performance tiers:
TierSummaryBest for
StandardSlower baseline tier with higher latency.General usage and non-latency-critical workflows.
PerformanceFastest tier with the lowest latency.Interactive workflows where speed is critical.