Skip to main content
The xysq Python SDK gives you direct API access to your memory vault — no MCP server required. Capture memories (including links, code snippets, and quoted passages — all tagged for source), surface relevant context, and build agents that remember across sessions.

​
Installation

pip install git+https://github.com/xysq-ai/sdk_python_xysq.git
Install optional extras for LiteLLM or Anthropic integrations: 
pip install 'xysq[agent] @ git+https://github.com/xysq-ai/sdk_python_xysq.git'
pip install 'xysq[claude] @ git+https://github.com/xysq-ai/sdk_python_xysq.git'
pip install 'xysq[all] @ git+https://github.com/xysq-ai/sdk_python_xysq.git'
Requires Python 3.11+

​
Get an API key

Get your API key at app.xysq.ai/connect. It starts with xysq_. Add it to a .env file in your project: 
XYSQ_API_KEY=xysq_...
Or pass it directly when constructing the client: 
from xysq import Xysq

client = Xysq(api_key="xysq_...")

​
Your first memory

from dotenv import load_dotenv
from xysq import Xysq

load_dotenv()

with Xysq() as client:
    # Store a memory
    client.memory.capture(
        content="User prefers type hints in all Python code",
        tags=["coding-style", "python"],
    )

    # Retrieve relevant memories
    memories = client.memory.surface("Python coding preferences")
    for m in memories:
        print(m.text)

    # Ask a question answered from memory
    result = client.memory.synthesize("What are the user's coding preferences?")
    print(result.answer)

​
Core concepts

The SDK exposes two namespaces on every client:
NamespacePurpose
client.memoryAll persistent memory — conversational facts, decisions, preferences, plus source-tagged links / quotes / code / chats
client.organiseFolders + uploaded files (PDFs, images, long documents)
The three core memory operations:
MethodWhat it does
capture(content)Store a memory permanently
surface(query)Retrieve the most relevant memories for a query
synthesize(query)Answer a natural language question from your memory bank
For external sources (URLs, quotes, code snippets, chat transcripts), use capture() with tags=["source:knowledge", "source_type:link"] (or quote/code/chat) and put type-specific fields (url, title, location, language) in metadata. They surface alongside regular memories in surface() and synthesize().

​
Source code

The SDK is open source on GitHub: xysq-ai/sdk_python_xysq Includes examples, community contributions, and the full source. Issues and PRs welcome.

​
What to explore next

Memory operations

Full reference for capture, surface, synthesize, list, delete, and status

Organise — folders & files

Upload PDFs, images, and long documents into a browsable folder tree

Team Vaults

Share memory across your entire team

XysqAgent & Integrations

Memory-aware agent wrapper with configurable context strategies and tool calling