mnemo-memory 0.1.1

Python SDK for Mnemo — AI memory that forgets intelligently

mnemoai

Python SDK for Mnemo — AI memory that forgets intelligently.

Installation

pip install mnemo-memory

Prerequisites

Start the Mnemo server:

npx @mnemoai/server
# or with OpenAI:
OPENAI_API_KEY=sk-... npx @mnemoai/server

Quick Start

from mnemo import MnemoClient

client = MnemoClient()  # connects to http://localhost:18100

# Store memories
client.store("User prefers dark mode", category="preference")
client.store("User is a backend engineer at Acme Corp", category="fact")

# Recall
results = client.recall("What does the user do?")
for memory in results:
    print(f"[{memory.score:.2f}] {memory.text}")

# Stats
stats = client.stats()
print(f"Total memories: {stats.totalEntries}")

# Delete
client.delete("memory-id")

# Cleanup
client.close()

Context Manager

with MnemoClient() as client:
    client.store("User likes Python")
    results = client.recall("programming language preferences")

Configuration

client = MnemoClient(
    base_url="http://localhost:8080",  # custom server URL
    timeout=60.0,                      # request timeout in seconds
)

API

Method	Description
client.store(text, category?, importance?, scope?)	Store a memory
client.recall(query, limit?, scope_filter?, category?)	Recall memories
client.delete(memory_id)	Delete a memory
client.stats()	Get statistics
client.health()	Health check

Types

from mnemo import Memory, RecallResult, Stats, MemoryCategory

• MemoryCategory: "preference" | "fact" | "decision" | "entity" | "other" | "reflection"
• Memory: text, score, category, importance, timestamp
• Stats: totalEntries, scopeCounts, categoryCounts

Links

• Documentation
• GitHub
• npm: @mnemoai/core
• npm: @mnemoai/server