Guides/Getting started

Getting started

Connect Claude to ContxOS: Add Persistent Memory as a Remote MCP Server

Add ContxOS to Claude.ai, Claude Code or any MCP client in about a minute: paste one URL, sign in with OAuth, and install the ContxOS commands.

Jose R. Gonzalez, CPA

Jose R. Gonzalez, CPA · Founder of ContxOS

Updated · 2 min read

In short

ContxOS is a remote MCP server at https://www.contxos.com/api/mcp. Add it as a custom connector in Claude.ai or Claude Code, leave the client ID and secret blank, and sign in. OAuth 2.0 with PKCE secures the connection, and there is nothing to install. Then tell the agent "update skills" and it installs the ContxOS commands (/contxos, /handoff, /finish, /workflow and others) so it knows how to use your knowledge base.

What you are connecting

ContxOS is a hosted knowledge base that your AI agent reads from and writes to. It is not a chat app of its own. Your agent in Claude.ai, Claude Code or another MCP client stays where it is and gains a memory that outlives every conversation.

The connection uses the Model Context Protocol (MCP), the open standard AI clients use to call external tools. ContxOS exposes a single tool to the agent, which keeps its tool list short and its context window free for your work.

Setup in Claude.ai

  1. Create a free ContxOS account at contxos.com. The free plan includes one project.
  2. In Claude.ai, open Settings → Connectors → Add custom connector.
  3. Paste https://www.contxos.com/api/mcp as the server URL. Leave Client ID and Client Secret blank.
  4. Click Add and sign in with your ContxOS account when prompted.
  5. In a new chat, type "sync chat" (or "update skills"). The agent installs the ContxOS instructions for chat use.

Setup in Claude Code

Claude Code connects to the same URL as a remote HTTP MCP server. Add it from your MCP settings or with the claude mcp add command, then authenticate when Claude Code asks.

Once connected, ask the agent to "sync ide". ContxOS writes its slash commands into your project’s .claude/commands folder: /contxos for search and routing, /handoff and /finish for closing sessions, /workflow for building procedures, and /dream for upkeep. The files are stamped with a version, and when ContxOS ships an update the agent tells you and offers a re-sync. It never re-syncs in the middle of a task.

Other MCP clients

Any client that supports remote MCP servers with OAuth sign-in can connect to the same URL. Claude.ai and Claude Code are what we use every day. Because the memory lives on ContxOS and not in any one client, what one agent records is there for the next agent, in whichever client it runs.

What to ask your agent

With ContxOS connected to Claude or another MCP client, requests like these work as written:

update skills
What projects do I have in ContxOS?
Search ContxOS for what we decided about the pricing page.
resume

Limits and requirements

  • Requires an MCP client that supports remote servers with OAuth. Claude.ai custom connectors and Claude Code both do.
  • The free plan includes one project and one seat. More projects and seats are on paid plans (see Pricing).
  • The installed commands are generated files. Edit your knowledge base, not the command files; a re-sync replaces hand edits.

Frequently asked questions

Do I need an API key?

No. You sign in with your ContxOS account through OAuth when you add the connector. There is no key to copy or rotate.

Does ContxOS read my Claude chat history?

No. ContxOS only holds what your agent writes to it, such as decisions, results, handoffs and workflows. It does not import transcripts.

Can I use it with more than one client at once?

Yes. Claude.ai, Claude Code and any other connected client read and write the same knowledge base.

Give your agent a memory that lasts

Connect ContxOS to Claude in about a minute. The free plan includes one project.