Getting Started with VEKTOR Slipstream
Welcome. This guide will get you from zero to a running VEKTOR memory layer in under 5 minutes.
Prerequisites
- Node.js 18+ installed
- Claude Desktop (for MCP integration) or any MCP-compatible agent
- A licence key from vektormemory.com/product
Step 1: Install
Step 2: Activate
The setup wizard will ask for your LLM provider (Groq, Gemini, OpenAI, Mistral, or Ollama) and API key. These power the AUDN curation and recall intelligence layers.
Step 3: Connect to Claude Desktop
The wizard will offer to auto-configure Claude Desktop MCP. Say yes. It writes the correct config to automatically.
Restart Claude Desktop. You should see VEKTOR tools available in your next conversation.
Step 4: Store your first memory
Or from Claude Desktop:
“Store a memory: I prefer concise responses with code examples”
Step 5: Recall
Or from Claude Desktop:
“What do you remember about my preferences?”
Useful Commands
| Command | Description |
|---|---|
| Store a memory | |
| Recall related memories | |
| Check health + memory count | |
| Run REM compression cycle | |
| Launch the D3 memory visualiser | |
| Get a summary of recent memories |
Troubleshooting
MCP tools not appearing in Claude Desktop?
Restart Claude Desktop fully. Check the config was written:
Memory count showing 0?
Run — check your DB path is correct and the SQLite file exists.
Embedding errors?
VEKTOR uses bundled ONNX models. No API key needed for embeddings. If you see errors, run again to re-run the setup wizard.
Stuck? Post in Usage & Troubleshooting with your version (), OS, and Node version.