Installation¶
Requirements¶
- Python 3.9 or higher
- An LLM API key (OpenAI, Azure, Anthropic, or any OpenAI-compatible provider)
Quick Install¶
In-Memory Only
The core package uses in-memory storage. Data is lost when your program exits.
Best for Most Users
Turso provides SQLite-based persistence that:
- Works offline
- Persists data to a local file
- Supports native vector search
- Requires no server setup
Package Options¶
| Package | Command | Use Case |
|---|---|---|
| Core | pip install agentic-graph-mem | Development, testing |
| Turso | pip install "agentic-graph-mem[libsql]" | Edge, offline, simple apps |
| Neo4j | pip install "agentic-graph-mem[neo4j]" | Graph-heavy workloads |
| Redis | pip install "agentic-graph-mem[redis]" | High-performance caching |
| All | pip install "agentic-graph-mem[all]" | Full enterprise stack |
Verify Installation¶
from graphmem import GraphMem, MemoryConfig
# Check version
import graphmem
print(f"GraphMem version: {graphmem.__version__}")
Next Steps¶
- Quick Start - Get up and running in 5 minutes
- Configuration - Configure GraphMem for your needs
- Storage Backends - Choose the right storage for your use case