AlgoDriven MCP server.

1. Click Add to Claude above. Claude opens with the AlgoDriven listing already selected.
2. Click Connect. A browser window opens — sign in with your AlgoDriven account and approve access.
3. In any chat, enable AlgoDriven from the connectors menu and ask away.

1. Open Customize → Connectors and click the + next to Connectors. From a chat you can also click +, hover Connectors and choose Manage connectors.
2. Search for AlgoDriven, select it and click Connect, then sign in as above.

An owner or admin has to add the connector for the organisation before staff can connect their own accounts.

1. Owner or admin: open Organization settings → Connectors and click Browse connectors.
2. Select AlgoDriven and click Add to your team.
3. Everyone else: open Customize → Connectors and click Connect next to AlgoDriven. Each person signs in with their own AlgoDriven account and sees only their own dealership's data.

If you'd rather not use the directory, add it as a custom connector instead: Customize → Connectors → + → Add custom connector, then paste the server URL.

https://mcp.algodriven.io/mcp

Add the server once, then sign in from your terminal.

1. Run the two commands below. User scope registers the server for every project on your machine.
2. claude mcp login opens your browser to sign in. On Claude Code versions before 2.1.186, start a session, run /mcp and choose algodriven → Authenticate instead.
3. Check it worked with claude mcp list — algodriven should come back connected.

# add the server (user scope — available in every project)
claude mcp add --transport http --scope user algodriven https://mcp.algodriven.io/mcp

# sign in with your AlgoDriven account
claude mcp login algodriven

ChatGPT reaches MCP servers through developer mode, which needs a paid ChatGPT plan. ChatGPT Pro is limited to read-and-fetch connectors, so the three retrieval tools work there but the two that start appraisals won't appear. Write support is in beta on ChatGPT Business, Enterprise and Edu, where an admin or owner turns developer mode on and publishes the app to the workspace.

1. Open Settings → Security and login and turn on Developer mode. In Business, Enterprise and Edu workspaces only admins and owners can turn it on, each for themselves — it isn't something an admin can switch on for your account.
2. Go to Settings → Plugins (or straight to chatgpt.com/plugins) and choose Create.
3. Name it AlgoDriven, paste the server URL below, and set authentication to OAuth.
4. Click Connect and sign in with your AlgoDriven account.
5. In a new chat, enable AlgoDriven from the plugins list.

https://mcp.algodriven.io/mcp

Microsoft's route is Copilot Studio. You add AlgoDriven as a tool on an agent, then publish that agent to Microsoft 365 Copilot, Teams or wherever your staff already work. Microsoft 365 Copilot itself has no add-a-server box like Claude or ChatGPT, so an admin or agent builder does this once for everyone. Copilot Studio supports the streamable HTTP transport, which is what our server uses.

1. In Copilot Studio, open the agent you want to use — or create one — and go to Tools → Add a tool → New tool → Model Context Protocol.
2. Enter AlgoDriven as the server name, a short description such as Drive, booking and appraisal records for the user's dealership, from DriveExpert and EvalExpert, and the server URL below. The agent uses the description to decide when to call the tool, so keep it plain.
3. Set authentication to OAuth 2.0 and the type to Dynamic discovery, then click Create. Copilot Studio finds our sign-in endpoints and registers itself, so there's no client ID or secret to enter.
4. On Add tool, choose Create a new connection, sign in with your AlgoDriven account and approve access, then click Add to agent.
5. Publish the agent. The five AlgoDriven tools show up under the agent's tools list.

https://mcp.algodriven.io/mcp

1. Choose whose login the agent uses. If the tool is set to use a shared connection, everyone who chats with the agent sees what the person who made the connection can see. To keep each person to their own dealership's data, have users sign in as themselves.
2. Check your data policies. Copilot Studio reaches MCP servers through Power Platform connectors, so a data loss prevention policy that restricts connectors will also restrict AlgoDriven until an admin allows it.
3. Prefer a connector file? You can instead create a custom connector in Power Apps from an OpenAPI file with the x-ms-agentic-protocol: mcp-streamable-1.0 extension on POST /mcp. The wizard above is Microsoft's recommended route. See Microsoft's guide for the details.

Gemini can reach remote MCP servers in three places, and they have different requirements. Pick the one that matches how you use Gemini.

Custom apps in the Gemini app are limited for now: you need to be 18 or over and in the US, signed in with a personal Google account (not a work or school one), with Keep Activity turned on, and with access to Gemini Spark. It's English only. Google's custom apps help page has the current rules.

1. On a computer, open gemini.google.com, click Settings at the bottom, then Connected Apps. If you don't see it, open Personal Intelligence first.
2. Under Custom apps, paste the server URL below into Add a custom app link and click Next.
3. Follow the prompts to sign in with your AlgoDriven account. Nothing more to enter — our server registers with Gemini automatically.
4. In a chat, type @ and choose AlgoDriven to make sure Gemini uses it. Once connected, it works in the Gemini web and mobile apps. You can only add it from the web app.

https://mcp.algodriven.io/mcp

For organisations on Gemini Enterprise, a team admin adds AlgoDriven once as a custom MCP server. Unlike the other assistants, Gemini Enterprise doesn't register itself: it asks for an OAuth client ID and secret, so contact us and we'll issue credentials for your organisation.

1. Admin: sign in to Gemini Enterprise, open Settings & help → Manage team → Connected apps and click Add MCP Server. (In the Google Cloud console the same option is a custom MCP server data store.)
2. Enter the server URL below and the name AlgoDriven, then choose OAuth 2.0 and fill in the values in the table.
3. Click Verify Auth, choose your data location, and add the connector. It starts switched off, so enable it to make it visible to your team.
4. Each person then authorises the connector once with their own AlgoDriven account.

https://mcp.algodriven.io/mcp

Google's terminal agent, for technical teams. Add the server, then sign in from inside a session.

1. Run the command below. User scope makes AlgoDriven available in every project.
2. Start gemini and run /mcp auth algodriven. Your browser opens — sign in with your AlgoDriven account.
3. Run /mcp to confirm algodriven shows as connected.

# add the server (user scope — available in every project)
gemini mcp add --transport http --scope user algodriven https://mcp.algodriven.io/mcp

One click if you have Cursor installed — or add the server from Cursor's settings.

1. Add to Cursor — Cursor opens and prompts you to install the server. To do it by hand, open Settings → Tools & MCP → New MCP Server (or edit ~/.cursor/mcp.json) and add the config below.
2. Quit and reopen Cursor so it registers the server.
3. Click Connect (shown as Needs login) next to algodriven. Your browser opens — sign in with your AlgoDriven account and approve access. The status dot turns green once it's ready.
4. Ask away in chat — Cursor picks up the AlgoDriven tools automatically.

{
  "mcpServers": {
    "algodriven": {
      "url": "https://mcp.algodriven.io/mcp"
    }
  }
}

Groq is for developers building their own agents: GroqCloud's OpenAI-compatible Responses API calls the AlgoDriven MCP server for you at inference time — Groq discovers the tools, runs the calls and feeds the results back to the model. There's no end-user connector UI, so authentication uses an access token in a header instead of a browser sign-in.

1. Get an AlgoDriven access token — contact us or ask your account manager, and we'll issue one for your integration.
2. Add the AlgoDriven server to the tools array of a Responses API request, as below.
3. That's it — Groq handles tool discovery and execution. See Groq's remote MCP docs for models and options.

curl https://api.groq.com/openai/v1/responses \
  -H "Authorization: Bearer $GROQ_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-oss-120b",
    "input": "How many test drives did we do last week?",
    "tools": [{
      "type": "mcp",
      "server_label": "algodriven",
      "server_url": "https://mcp.algodriven.io/mcp",
      "headers": { "Authorization": "Bearer YOUR_ALGODRIVEN_TOKEN" },
      "require_approval": "never"
    }]
  }'

Any client that supports remote MCP servers over HTTP with OAuth — VS Code, Windsurf, Zed and others. Most read a config file in this shape:

{
  "mcpServers": {
    "algodriven": {
      "type": "http",
      "url": "https://mcp.algodriven.io/mcp"
    }
  }
}

The exact file location and key name vary by client — check your client's MCP documentation. Pick the streamable HTTP transport rather than SSE, which most clients now treat as deprecated. On first use it will open a browser for you to sign in with your AlgoDriven account.