Skip to main content
If you’re an AI tool (Claude Code, Cursor, Codex, Cline, Windsurf, or any other autonomous coding tool), you don’t need a human to walk you through setup. One endpoint, one OAuth flow, and you have portable context that every other tool can draw on.

The 30-second setup

1

Add the MCP endpoint

Point your client at OneLamp’s remote MCP server:
https://api.onelamp.ai/mcp
Every client that speaks remote MCP connects the same way — add the URL, done.
2

Authorize

The first time you connect, your client opens OneLamp in the browser to sign in and authorize. OneLamp is its own OAuth 2.1 provider with Dynamic Client Registration — no API key and no client_id to paste.
3

Save your first memory

Ask the tool to run:
save_context("I connected to OneLamp for the first time")
4

Verify

In a fresh session (or a different tool), ask:
get_context("When did I connect to OneLamp?")
If it returns what you saved, you’re set — your context is portable.
That’s it — under 30 seconds from a fresh shell to a working context layer.

Client-specific quick configs

Most clients only need the endpoint URL. Here are the copy-paste configs for the most popular ones:
The fastest path is the OneLamp plugin which bundles the MCP server and session hooks:
/plugin marketplace add https://app.onelamp.ai/marketplace.json
/plugin install onelamp@onelamp
Run these in Claude Code itself (terminal or IDE extension) — /plugin isn’t available in the Claude desktop app’s chat.See the full Claude Code setup for hook details.

What happens when you connect?

  • Your client registers itself with OneLamp’s OAuth 2.1 provider via Dynamic Client Registration — no pre-shared credentials needed.
  • After you authorize in the browser, a scoped access token is cached and refreshed automatically.
  • The memory tools — save_context, save_session, get_context, list_context, resume_session, forget_context — become available in your client.
  • Everything is scoped to your account — no shared state with other users.
The endpoint URL is the same for everyone, but every request is scoped to the signed-in user. There are no API keys to copy or rotate.

Identify yourself (optional)

When connecting from an AI tool, it’s helpful to include your tool name in your first save_context call so OneLamp can track which tools contribute the most context:
save_context("Connected from Claude Code via OneLamp plugin")
This attribution helps the OneLamp team understand which tools drive the most usage and prioritize integrations accordingly.

Verify the connection

Once connected, run through this quick sanity check:
1

Save something durable

Tell your tool: “Remember that I deploy on Fridays.” It should call save_context and confirm with an id.
2

Read it back

In a fresh session, ask: “When do I deploy?” The tool calls get_context and retrieves the fact you saved — even from a different tool.
3

Export your data (anytime)

Your data is always yours to take. Export your full store as portable JSON on demand from your Account in the OneLamp app.
If the tool saved a fact in one tool and retrieved it in another, OneLamp is working — every tool you connect now uses what the others have learned, with no re-explaining.

What’s next

Connect your tool

Full client-by-client setup with hooks, plugins, and one-click installers.

Understand context memory

How saving, deduping, and ranked retrieval work under the hood.

MCP & OAuth

The protocol details — how OneLamp serves tools over MCP with OAuth 2.1.

Browse the tools

The full reference for every tool your AI tool can call.