Skip to main content

​
What is Velixar?

Velixar is cognitive memory for AI. Not a vector database — a reasoning layer that gives your AI applications persistent memory, a knowledge graph, identity awareness, contradiction detection, and belief tracking across every session.

Quickstart

Store your first memory in under 2 minutes

API Reference

Memory, graph, identity, and webhook endpoints

MCP Server

25 cognitive tools for Claude, Kiro, Cursor

Python SDK

pip install velixar

JavaScript SDK

npm install velixar

Knowledge Graph

Entities, relationships, traversal

CI/CD Webhook

Store deployments and decisions as memories

Security

Workspace isolation and content scanning

​
Why Velixar?

LLMs are stateless. Every conversation starts from zero. Most “memory” solutions just add a vector store. Velixar goes further.
CapabilityBasic memory toolsVelixar
Store and search✅✅
Workspace isolation❌✅ Per-project scoping, zero bleed
Knowledge graph❌✅ Entities and relationships extracted automatically
Contradiction detection❌✅ Surface conflicting facts with resolution guidance
Belief evolution❌✅ Track how understanding changes over time
Identity model❌✅ Persistent user profile — preferences, expertise, goals
Session continuity❌✅ Save and restore full session context
CI/CD integration❌✅ Webhook + GitHub Actions for deployment and decision memory
Cognitive prompts❌✅ 16 guided reasoning workflows

​
How It Works

1

Store

Send memories to the API — facts, preferences, decisions, session summaries. Each memory is embedded, indexed, and entities are extracted into a knowledge graph.
2

Reason

Your AI doesn’t just retrieve text — it orients itself in the workspace, traverses entity relationships, detects contradictions, and tracks how beliefs evolved.
3

Recall

Query with natural language. Velixar returns semantically relevant memories ranked by relevance, with full provenance and confidence signals.

​
The MCP Server

The fastest way to use Velixar is the MCP server — one install gives any compatible AI assistant 25 cognitive tools, 5 live resources, and 16 workflow prompts: 
npm install -g velixar-mcp-server
Your assistant immediately gets orientation, contradiction detection, knowledge graph traversal, identity awareness, session continuity, and import/export — all through the standard MCP protocol. Full MCP server documentation →

​
SDKs

Use Velixar directly from code: 
# Python
pip install velixar

# JavaScript / TypeScript
npm install velixar
Both SDKs support memory CRUD, knowledge graph traversal, identity, and the CI/CD webhook. Python includes both sync and async clients.

​
Key Concepts

  • Memories — The atomic unit of knowledge. Text with tags, tiers, and provenance.
  • Tiers — Pinned (0), Session (1), Semantic (2), Organization (3). Different persistence and visibility.
  • Knowledge Graph — Entities and relationships extracted automatically from memories.
  • Workspace Isolation — Every memory is scoped to a workspace. Projects never bleed.
  • Org Memory — Share knowledge across teams with approval workflows.