Quick Start

Get Synapse running and connected to an MCP-compatible client in minutes.

1. Install

npm install -g synapse-cortex

:::info Requires Node.js >= 22.6.0. Check your version with node --version. :::

2. Initialize Your Workspace

Navigate to your project root and run setup:

synapse setup
synapse doctor

setup creates the local SQLite database and installs native dependencies. doctor confirms the runtime is healthy.

3. Connect Your MCP Client

Add Synapse as an MCP server in your client configuration:

{
  "mcpServers": {
    "synapse": {
      "command": "synapse",
      "startup_timeout_sec": 30,
      "env": {
        "MCP_MODE": "stdio",
        "SYNAPSE_CONFIG": "~/.synapse/config/synapse.config.json",
        "SYNAPSE_INDEX_BACKEND": "sqlite-vec",
        "SYNAPSE_MEMORY_ENABLED": "true"
      }
    }
  }
}

This works with Claude Desktop, Claude Code, Cursor, Windsurf, Cline, Continue, Gemini CLI, and other MCP-compatible clients.

4. Query Your Project

Once connected, your client can call Synapse tools for code search, memory, and graph queries. Start with an architecture overview:

"Using Synapse, analyze the project structure and identify the core services and their relationships."

Synapse can use summarize_project, search_hybrid, and project_tree to produce a semantically accurate map of your codebase.

5. Store Project Knowledge

Store a project-level fact that persists across future sessions:

"Remember that the authentication module uses JWT with 15-minute expiry. Store this in memory."

Synapse stores this with memory_store, and the fact can be retrieved later with memory_recall.

What's Next?

• Understand Code Intelligence: semantic search, symbol finding, and AST-aware retrieval.
• Explore the Knowledge Graph: temporal triples and architectural decision tracking.
• Learn about Persistent Memory: cross-session recall and scoped isolation.