Integrations

One memory.
Every agent.

Korely exposes a REST API, a Python and Node SDK, and a CLI. Any agent, workflow engine, or custom script that can make an HTTP request gets the same typed, bi-temporal memory store.

Workflow automation

n8n

Add Korely as an HTTP node in your n8n workflows. Write memories from triggers, query the graph from conditions.

Setup guide

LangChain agent framework

LangGraph

Wrap Korely memory tools as LangGraph nodes. Give stateful graphs a long-term store that outlives the run.

Setup guide

TypeScript / Next.js

Vercel AI SDK

Call the REST API or the Node SDK inside your AI SDK routes. Memory that persists across user sessions out of the box.

Setup guide

Open-source agent runtime

OpenClaw

OpenClaw works with Korely today via the REST API. Store and retrieve typed facts alongside your tool calls.

Setup guide

One API

Any language. Any framework.

Korely Agent Memory is a plain REST API hosted in Helsinki. One Authorization: Bearer kor_live_… header is all the auth you need. No OAuth dance, no SDK required to get started.

If you prefer a typed client, pip install korely-memory or npm install korely-memory gives you Python and Node SDKs that mirror the REST surface method-for-method. The Python package also ships the korely CLI for interactive use and shell scripts.

Because the integration point is HTTP, the same memory layer works whether your agent runs in Python, TypeScript, n8n, or a custom runtime. See the API reference for full endpoint schemas.

Don't see your stack?

If it can make an HTTP call, it works.

Tell us what you're building and we'll add a setup guide. The REST API is the integration layer — nothing proprietary to install.

[email protected]