Skip to content

getting-started.md

Latest commit

 

History

History
189 lines (129 loc) · 4.08 KB

getting-started.md

File metadata and controls

189 lines (129 loc) · 4.08 KB

Getting Started

CodePrysm is a powerful tool for analyzing code repositories and generating relationship graphs using Tree-sitter abstract syntax trees. This system transforms source code into a searchable knowledge graph, enabling semantic code search, dependency analysis, and intelligent code navigation.

Table of Contents

  1. Install Prerequisites
  2. Install CodePrysm
  3. Start Qdrant
  4. Run Initial Indexing
  5. Setup MCP Server
  6. Verify Installation

For Docker-based setup, see the Docker Setup Guide.

1. Install Prerequisites

  • Rust 1.85+: Install Rust
  • Docker: Required for running Qdrant vector database

Optional:

  • just: Command runner for development shortcuts

2. Install CodePrysm

From crates.io (Recommended)

cargo install codeprysm-cli

From Source

git clone https://github.com/codeprysm/codeprysm.git
cd codeprysm
cargo build --release

The binary will be available at ./target/release/codeprysm.

GPU Acceleration (Optional)

For faster embedding generation, build with GPU support:

# macOS (Apple Silicon)
cargo install codeprysm-cli --features metal

# Linux (NVIDIA)
cargo install codeprysm-cli --features cuda

3. Start Qdrant

CodePrysm uses Qdrant for semantic search. Start it with Docker:

docker run -d --name qdrant \
  -p 6333:6333 -p 6334:6334 \
  -v qdrant_storage:/qdrant/storage \
  qdrant/qdrant:latest

Or using just:

just qdrant-start

4. Run Initial Indexing

Navigate to your repository and run the initial indexing:

cd /path/to/your/repo
codeprysm init

This creates a .codeprysm/ directory containing:

  • Code graph with all entities and relationships
  • Merkle tree for incremental updates
  • Semantic embeddings indexed in Qdrant

The .codeprysm/ directory is automatically added to .gitignore.

Incremental Updates

After the initial indexing, use update for faster incremental updates:

codeprysm update

5. Setup MCP Server

Option A: VS Code Integration

Create .vscode/mcp.json in your repository:

{
    "servers": {
        "codeprysm": {
            "type": "stdio",
            "command": "codeprysm",
            "args": ["mcp"]
        }
    }
}

Option B: Claude Desktop Integration

Add to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "codeprysm": {
      "command": "codeprysm",
      "args": ["mcp", "--root", "/path/to/your/repo"]
    }
  }
}

Option C: Manual MCP Server

Start the MCP server manually:

codeprysm mcp --root /path/to/your/repo --qdrant-url http://localhost:6334

6. Verify Installation

Check Qdrant Connection

codeprysm stats

Test Search

codeprysm search "function that handles authentication"

In VS Code

  1. Open the Chat panel
  2. Select 'Agent' chat mode
  3. Test the search: #search_graph_nodes authentication functions

You should see relevant code snippets from your codebase.

Troubleshooting

Qdrant Connection Issues

Ensure Qdrant is running:

docker ps | grep qdrant

Check Qdrant health:

curl http://localhost:6333/health

Slow Indexing

  • Enable GPU acceleration with --features metal (macOS) or --features cuda (Linux)
  • For large repositories (>10K files), initial indexing may take 5-20 minutes

Memory Issues

For very large repositories:

  • Ensure sufficient RAM (4GB+ for medium repos, 16GB+ for large repos)
  • Use incremental updates (codeprysm update) instead of full rebuilds

Next Steps