cubememory 0.1.2

Official Python SDK for Cube Memory — semantic search with 100% exact recall

cubememory

Official Python SDK for Cube Memory — semantic search with 100% exact recall, isolated multi-tenant islands, and native MCP support.

Install

pip install cubememory          # sync (uses requests)
pip install cubememory[async]   # + async support (uses httpx)

Quick Start

from cubememory import CubeClient

cube = CubeClient(
    base_url="https://cubememory.com.br/gateway",
    api_key="cm_live_YOUR_KEY",
    project_id="proj_YOUR_PROJECT",
)

# Semantic search — 100% exact cosine recall (no ANN approximation)
results = cube.index("my-docs").search("wireless headphones", limit=10)
for r in results:
    print(f"[{r.score:.2f}] {r.text_preview}")

# Index a document
cube.index("my-docs").insert(
    id="doc-001",
    text="Noise cancelling Sony WH-1000XM5",
    metadata={"price": 349, "brand": "Sony"},
)

# LLM Memory (agent remembers between sessions)
cube.memory.store("User prefers dark mode and speaks Portuguese")
ctx = cube.memory.search("user preferences")

# PKM Notes
cube.notes.upsert(
    title="Meeting Q3",
    text="Discussed the roadmap and timeline.",
    tags=["meeting", "q3"],
)
notes = cube.notes.search("roadmap")

Async Usage

import asyncio
from cubememory.client import AsyncCubeClient

async def main():
    cube = AsyncCubeClient(
        base_url="https://cubememory.com.br/gateway",
        api_key="cm_live_YOUR_KEY",
        project_id="proj_YOUR_PROJECT",
    )
    results = await cube.index("my-docs").search("wireless headphones")
    memories = await cube.memory.search("user preferences")

asyncio.run(main())

API Reference

CubeClient(base_url, api_key, project_id, timeout=30.0)

Param	Type	Description
base_url	str	Gateway URL, e.g. https://cubememory.com.br/gateway
api_key	str	Your API key (cm_live_xxx)
project_id	str	Your project ID (proj_xxx)

---

cube.index(name)

Method	Description
.insert(id, text, metadata?)	Index a document
.insert_many(docs[])	Batch index
.search(query, limit?, threshold?, filter?)	Semantic search → List[SearchResult]
.get(id)	Fetch by ID
.delete(id)	Remove document
.stats()	Index statistics

SearchResult fields: id, score, text_preview, metadata

---

cube.memory

Agent memory — persists LLM context between sessions.

Method	Description
.store(text, layer?)	Save memory ("long-term" or "short-term")
.search(query, limit?)	Semantic search → List[SearchResult]
.list(limit?)	List all memories
.delete(id)	Delete a memory

---

cube.notes

PKM Notes — Obsidian-style knowledge base with AI auto-connections.

Method	Description
.upsert(title, text, id?, tags?)	Create or update a note
.list()	List all notes → List[Note]
.get(id)	Full note with related notes → NoteDetail
.search(query, limit?)	Semantic search → List[Note]
.delete(id)	Delete note and its chunks

---

MCP Integration

Connect Claude, GPT, Cursor, or any MCP-compatible LLM:

{
  "mcpServers": {
    "cube-memory": {
      "type": "http",
      "url": "https://cubememory.com.br/gateway/v1/mcp",
      "headers": {
        "Authorization": "Bearer cm_live_YOUR_KEY",
        "X-Project-Id": "proj_YOUR_PROJECT"
      }
    }
  }
}

Available MCP tools: search_memory, store_memory, list_memories, search_notes, get_note.

---

LangChain / Agent Integration

from cubememory import CubeClient

cube = CubeClient(base_url, api_key, project_id)

# Use as a LangChain retriever-style tool
def retrieve(query: str) -> list[str]:
    results = cube.index("knowledge-base").search(query, limit=5)
    return [r.text_preview for r in results]

# Agent memory tool
def remember(text: str) -> str:
    cube.memory.store(text)
    return f"Stored: {text[:60]}..."

def recall(query: str) -> str:
    results = cube.memory.search(query, limit=3)
    return "\n".join(r.text_preview for r in results)

---

Why Cube Memory?

	Cube Memory	Bancos vetoriais baseados em ANN/HNSW
100% exact recall	✅ cosine float32	❌ ANN approximation
Multi-tenant isolation	✅ island per project	⚠️ shared index with filters
MCP native	✅ built-in	❌
Self-host footprint	✅ minimal, single-host friendly	❌ heavy multi-service clusters
LGPD/GDPR	✅ data never leaves your server	❌ cloud by default

---

Links

• Dashboard: cubememory.com.br
• npm SDK: @cubememory/sdk
• Issues: github.com/cubememory/sdk-python