# VEKTOR Memory — Persistent Memory for AI Agents
https://vektormemory.com

## What VEKTOR Does
VEKTOR is a local-first persistent memory SDK for AI agents. It provides 8ms recall latency, zero cloud dependency, and automatic graph wiring across semantic, causal, temporal, and entity connections via the MAGMA architecture.

## WebMCP Tools Available
This site exposes 7 WebMCP tools at /.well-known/webmcp.json for AI agents using Chrome 146+ or WebMCP-compatible browsers:

1. store_memory — POST /api/memory/store — Save facts to the VEKTOR graph (requires X-VEKTOR-License header)
2. query_memory — POST /api/memory/query — Natural language search with 8ms avg recall
3. memory_status — GET /api/memory/status — Real-time system health (no license required)
4. activate_vektor_license — POST /api/license/activate — Activate a VEKTOR Pro license key
5. request_vektor_demo — POST /api/demo/request — Schedule a personalised demo
6. compare_vektor — POST /api/compare — Structured comparison vs competitors (mem0, openai, etc.)
7. vektor_agent — POST /api/webmcp/agent — Autonomous reasoning over the MAGMA causal graph

## Key Features
- MAGMA Graph Architecture: 4-layer memory (semantic + temporal + causal + entity)
- 8ms Recall Latency: Local SQLite-vec, no network round-trip
- REM Dream Cycle: Automatic compression at 50:1 ratio during idle
- Zero Cloud Dependency: All data stays on your machine
- LoCoMo Benchmark: 76.8% accuracy (v1.6.0), beating Mem0's published 62.47%
- WebMCP Native: First memory SDK with full WebMCP implementation

## Getting Started
- Docs: https://vektormemory.com/docs
- Product: https://vektormemory.com/product
- Compare: https://vektormemory.com/compare
- Downloads: https://vektormemory.com/downloads

## Pricing
- Free: Limited temporary memory (non-persistent)
- Pro: $9/month — Persistent storage, REM cycle, multi-agent, MAGMA graph, WebMCP access

## Install
npm install -g vektor-slipstream

## Contact
hello@vektormemory.com