Setup guide · 2 minutes

Connect your AI assistant
to Ovrwhlm.

Connect your AI assistant once, then capture, sort, and triage your tasks with the Eisenhower Matrix — just by talking.

How this works

Your AI assistant is who you talk to. Ask it to capture, sort, and triage your tasks and it does the work in Ovrwhlm. Open the Ovrwhlm app on your phone any time to see your matrix and backlog.

Before you start

Ovrwhlm is a standard remote MCP server, so it works with any assistant that supports MCP — ChatGPT, Claude, Gemini, Perplexity, Cursor, VS Code and more. You need an Ovrwhlm account (the same email you use in the Ovrwhlm app). Some assistants only allow adding connectors on desktop or the web, and some plans limit custom connectors.

Connect in 4 steps

  1. Pick your assistant

    Any assistant that supports remote MCP servers works. Find yours in the list below for exact steps.

  2. Add Ovrwhlm as a connector

    In your assistant, add a custom connector or remote MCP server. The only field that matters is the server URL — paste Ovrwhlm’s address below.

    https://ovrwhlm.kecker.co/mcp
  3. Sign in to Ovrwhlm

    An Ovrwhlm window opens. Sign in with your email and approve access. You only do this once. No token or config file to edit — sign-in handles it.

  4. You’re connected

    Your assistant can now manage your tasks. Try one of the prompts below.

Steps for your assistant

These are the popular ones. Ovrwhlm is a standard MCP server, so if yours isn’t listed, add the same URL as a remote MCP server. Menu names and plan limits change, so look for the closest match.

Chat apps

ChatGPT
  1. In ChatGPT open Settings → Connectors → Advanced and turn on Developer mode. It is limited to some plans, and on Business and Enterprise an admin may need to allow it first.
  2. Back in Connectors choose Create. Name it, paste the server URL, and set authentication to OAuth.
  3. Start a chat, open the + menu, choose Developer mode and switch the connector on. Sign in to Ovrwhlm when asked.
Claude (web, Desktop, mobile)
  1. Use claude.ai in a browser or Claude Desktop. Custom connectors can’t be added from the mobile app, but once added they work there too.
  2. Open Settings → Connectors (newer versions label it Customize → Connectors) and choose Add custom connector.
  3. Paste the server URL and add it. Sign in to Ovrwhlm when the window opens, then switch the connector on in a chat.

The free plan allows one custom connector. On Team and Enterprise an organization owner adds it first, then each member clicks Connect.

Perplexity
  1. Open Settings → Connectors and choose + Custom connector, then Remote. It needs a paid plan, and on Enterprise an admin must allow it.
  2. Enter a name and the server URL, choose Streamable HTTP if asked, and set authentication to OAuth.
  3. Finish the Ovrwhlm sign-in in the browser tab, then enable the connector in a thread.
Mistral Le Chat
  1. In Le Chat open the side panel, then Intelligence → Connectors, and choose + Add Connector.
  2. Switch to the Custom MCP Connector tab, enter a name and the server URL, and choose Connect.
  3. Sign in to Ovrwhlm, then enable the connector in a new chat. Adding one needs an admin on team workspaces.

Coding tools and CLIs

Claude Code
claude mcp add --transport http ovrwhlm https://ovrwhlm.kecker.co/mcp

Then run /mcp inside a session and choose Authenticate (or run claude mcp login ovrwhlm in your shell) to sign in to Ovrwhlm. Check with claude mcp list.

OpenAI Codex CLI

Add this to ~/.codex/config.toml:

[mcp_servers.ovrwhlm]
url = "https://ovrwhlm.kecker.co/mcp"

Then run codex mcp login ovrwhlm to sign in to Ovrwhlm. codex mcp list shows what is configured.

Gemini CLI
gemini mcp add -s user --transport http ovrwhlm https://ovrwhlm.kecker.co/mcp

Restart Gemini CLI and run /mcp to check the connection. If it asks you to sign in, follow the browser prompt. Prefer editing by hand? Put "ovrwhlm": { "httpUrl": "https://ovrwhlm.kecker.co/mcp" } under mcpServers in ~/.gemini/settings.json.

Google Antigravity CLI (agy)
agy mcp add ovrwhlm https://ovrwhlm.kecker.co/mcp

The address is recognised as an HTTP server automatically. Run agy mcp list to confirm, then sign in to Ovrwhlm if prompted.

Cursor

Add this to ~/.cursor/mcp.json (or .cursor/mcp.json inside one project):

{
  "mcpServers": {
    "ovrwhlm": { "url": "https://ovrwhlm.kecker.co/mcp" }
  }
}

Cursor opens a browser tab to sign in to Ovrwhlm the first time a tool is used. Restart Cursor if the server doesn’t show up.

VS Code (GitHub Copilot)

Press Cmd/Ctrl+Shift+P, run MCP: Add Server, pick HTTP and paste the URL. Or add this to .vscode/mcp.json:

{
  "servers": {
    "ovrwhlm": { "type": "http", "url": "https://ovrwhlm.kecker.co/mcp" }
  }
}

Note the key is servers, not mcpServers. Use Copilot Chat in Agent mode, trust the server, and sign in to Ovrwhlm when the browser opens.

Windsurf

Open the MCP config from the Cascade panel (… menu → MCP config) and add:

{
  "mcpServers": {
    "ovrwhlm": { "serverUrl": "https://ovrwhlm.kecker.co/mcp" }
  }
}

Refresh the MCPs list, or restart Windsurf, then sign in to Ovrwhlm when prompted.

Zed

Open Settings → AI → MCP Servers, or add this to settings.json:

{
  "context_servers": {
    "ovrwhlm": { "url": "https://ovrwhlm.kecker.co/mcp" }
  }
}

Zed starts the sign-in to Ovrwhlm for you. The key must be context_servers.

Any other MCP client

Add a remote MCP server (Streamable HTTP) with the URL above. Sign-in uses standard OAuth, so there is no token or key to copy. If your client only supports local (stdio) servers, bridge it with npx mcp-remote https://ovrwhlm.kecker.co/mcp.

Try saying

If something’s not working

My assistant says it can’t find my tasks
Open your assistant’s connector settings, find Ovrwhlm, and confirm it shows your email (signed in). If not, open it and sign in again.
I don’t see a way to add a custom connector
Many assistants only allow it on desktop or the web, not in their mobile app, and some limit it by plan or need an admin to allow it. Set it up there once and it usually carries over to mobile.
My assistant only supports local (stdio) servers
Bridge it with npx mcp-remote https://ovrwhlm.kecker.co/mcp, or use a client that supports remote MCP servers directly.
The old guide told me to edit a JSON config file
That’s no longer needed for assistants that support remote connectors. Just add the URL above and sign in — no token to copy.