Synapps

Code intelligence MCP server — gives AI coding agents a queryable graph of your codebase.

What it does

• Indexes C#, Python, TypeScript/JavaScript, and Java codebases using tree-sitter and Language Server Protocol
• Stores symbols, call chains, and relationships in a Memgraph graph database
• Exposes semantic query tools as MCP tools — AI agents can find usages, trace call trees, and analyze change impact without reading every file
• Supports incremental sync via git-diff-based change detection and live file watching

Install

pip install synapps-mcp
docker compose up -d

The docker compose up -d command starts a local Memgraph instance (required for the graph database).

Configure your MCP client

Add Synapps to your MCP client configuration (e.g., Claude Code ~/.claude.json):

{
  "mcpServers": {
    "synapps": {
      "command": "synapps-mcp"
    }
  }
}

Then index your project:

synapps index /path/to/your/project

Supported Languages

• C#
• Python
• TypeScript / JavaScript
• Java

Key MCP Tools

• get_context_for — full context for a symbol before editing (callers, callees, tests); use scope="impact" for change impact analysis
• find_usages — who calls a method or uses any symbol across the codebase
• find_callees — what a method depends on downstream
• find_entry_points — find API/controller entry points reaching a method
• search_symbols — find symbols by name, kind, file, or namespace

Links

• GitHub
• Changelog