Skip to main content

Agent with Persistent Memory

Build an agent that stores and recalls information across sessions using ZeroMemory.

Prerequisites​

Store Memories​

import requests

TOKEN = "your-api-key"
BASE = "https://api.ainative.studio/api/v1/public/memory/v2"
HEADERS = {"Authorization": f"Bearer {TOKEN}", "Content-Type": "application/json"}

# Store user preferences
requests.post(f"{BASE}/remember", headers=HEADERS, json={
    "content": "User prefers dark mode and uses Python for backend",
    "entity_id": "user_123",
    "memory_type": "semantic",
    "importance": 0.9,
    "tags": ["preferences"],
})

# Store a conversation fact
requests.post(f"{BASE}/remember", headers=HEADERS, json={
    "content": "User is building a RAG chatbot for their company wiki",
    "entity_id": "user_123",
    "memory_type": "episodic",
    "importance": 0.7,
    "tags": ["project"],
})

# Store a relationship
requests.post(f"{BASE}/relate", headers=HEADERS, json={
    "subject": "user_123",
    "predicate": "works_on",
    "object": "RAG Chatbot Project",
    "confidence": 0.9,
})

Recall in a New Session​

# Later, in a completely new session...
response = requests.post(f"{BASE}/recall", headers=HEADERS, json={
    "query": "What is the user working on?",
    "entity_id": "user_123",
    "limit": 5,
})

for memory in response.json()["results"]:
    print(f"[{memory['score']:.2f}] {memory['content']}")

Output:

[0.94] User is building a RAG chatbot for their company wiki
[0.72] User prefers dark mode and uses Python for backend

Build a Profile​

profile = requests.get(
    f"{BASE}/profile/user_123",
    headers=HEADERS,
).json()

print(profile)

What to Try Next​

  • Use GraphRAG for relationship-aware retrieval
  • Add reflection for AI-generated insights
  • Connect via MCP for agent tool access