kortexio-contextmemory

Thin header helper for ContextMemory — the OpenAI-compatible agentic memory gateway behind Kortexio.

This is not a full SDK. It only builds the auth / tenant headers (Authorization, X-App-Id, X-User-Id, X-Session-Id) so you can keep using the official OpenAI Python client (or httpx / requests).

Beta (0.0.1b*). APIs may still change.
PyPI name: kortexio-contextmemory · Import name: contextmemory

Install

pip install kortexio-contextmemory openai

Quick start (OpenAI Python client)

Point the client at your gateway /v1 base URL (self-host or Cloud):

import os
from openai import OpenAI
from contextmemory import openai_client_kwargs

client = OpenAI(
    **openai_client_kwargs(
        base_url=os.environ.get("CONTEXTMEMORY_BASE_URL", "http://localhost:5100/v1"),
        api_key=os.environ["CONTEXTMEMORY_API_KEY"],  # cm_live_… or self-host key
        app_id="demo-dev",
        user_id="user-42",
        session_id="sess-abc",
    )
)

completion = client.chat.completions.create(
    model="qwen3.5:9b",
    messages=[{"role": "user", "content": "Remember that my staging DB is postgres-staging."}],
)
print(completion.choices[0].message.content)

Headers only (httpx, requests)

import httpx
from contextmemory import headers

h = headers(
    api_key="cm_live_...",
    app_id="demo-dev",
    user_id="user-42",
    session_id="sess-abc",
)

r = httpx.post(
    "http://localhost:5100/v1/chat/completions",
    headers=h,
    json={
        "model": "qwen3.5:9b",
        "messages": [{"role": "user", "content": "Hello"}],
    },
)
print(r.json())

API

Required: api_key (and base_url for the OpenAI helper)
Optional: app_id, user_id, session_id, extra (merged into headers)

Cloud vs self-host

Gateway docs: GitHub README · compare (why not RAG)

License

AGPL-3.0-or-later — same as the ContextMemory core.