Use Blevo in any AI assistant.

Until Blevo is listed in ChatGPT’s plugin directory, add it yourself with developer mode. OpenAI is still rolling these menus out, so names can differ slightly on your account.

1. On chatgpt.com, open Settings›Security and login and turn on Developer mode. (On some accounts it’s under Settings›Apps›Advanced settings; on Business, Enterprise and Edu an admin may need to allow it first.)
2. Open Plugins in the sidebar, choose + and create an app for a remote MCP server. Name it Blevo, paste https://mcp.blevo.app/mcp and pick OAuth.
3. A Blevo page opens: sign in with the account you use in the app (or enter the code from You›Settings›AI›Connect AI assistants in the app), choose what to share, and approve.
4. In a chat, choose Developer mode from the + menu, switch Blevo on, and ask “How did my week go?”

OpenAI’s help centre says developer-mode apps on Pro are limited to reading; changes (like setting a goal) are for Business, Enterprise and Edu. Reading your data works everywhere developer mode does.

If Blevo shows up under Plugins

Once it’s listed, you won’t need developer mode: open Settings›Plugins (or Plugins in the sidebar), pick Blevo, choose Install plugin, then Connect and sign in.

Dots (always-on agents)

A dot uses the same plugin connections as the rest of ChatGPT, so once Blevo is connected your dot can use it — for example to check your week and nudge you. OpenAI says dots use only read-only tools when they research on their own. Manage access in the Plugins tab, or on mobile in the dot’s profile under Customize›Plugins. Dots are rolling out to Pro and Business Premium and aren’t available to under-18s.

1. On claude.ai or in Claude Desktop, open Customize›Connectors.
2. Click +, then Add custom connector. Name it Blevo, paste https://mcp.blevo.app/mcp and click Add. Leave the advanced OAuth fields empty.
3. Click Connect. A Blevo page opens: sign in with the account you use in the app (or enter the code from You›Settings›AI›Connect AI assistants in the app), choose what to share, and approve.
4. In a chat, open +›Connectors, switch Blevo on and ask something like “Should I train hard or rest today?”

Team and Enterprise: an Owner adds it once in Organization settings›Connectors → Add → Custom → Web. Everyone else then finds Blevo in Customize›Connectors and clicks Connect to link their own Blevo account.

Mobile: add Blevo on the web or in Claude Desktop first; it’s then available in Claude for iOS and Android when you’re signed in to the same account.

Claude Code

claude mcp add --transport http blevo https://mcp.blevo.app/mcp

Then run /mcp inside Claude Code, pick blevo and sign in in your browser. Connectors you added on claude.ai are also available in Claude Code when you’re signed in with the same account.

On grok.com

1. Go to grok.com/connectors (the Connectors page).
2. Choose New Connector, then Custom, and paste https://mcp.blevo.app/mcp.
3. A Blevo page opens: sign in with the account you use in the app (or enter the code from You›Settings›AI›Connect AI assistants in the app), choose what to share, and approve.
4. Ask Grok about your training — it can use Blevo in any chat.

On Grok Business and Enterprise, a team admin makes connectors available first.

Grok Bot

Grok Bot is in beta and doesn’t have a form for your own servers yet. Ask it in a chat: “Add this MCP server: https://mcp.blevo.app/mcp”. Confirm the details it shows, sign in to Blevo when your browser opens, then type @ in a chat to use it. If Blevo appears in Grok Bot’s Marketplace, choose Add there instead.

Meta’s assistants handle connectors differently from the others, and the details are still changing:

• Muse: custom connectors are created in a chat. Ask Muse to “create a custom connector for Blevo” and give it the connector URL https://mcp.blevo.app/mcp (a remote MCP server that signs in with OAuth). Muse walks you through it; you then sign in to Blevo and approve. Meta doesn’t review custom connectors.
• Meta AI app, web and glasses: connectors there come from Meta’s connector programme, which is in developer preview. If Blevo appears in the connector list in settings, connect it from there.

Either way, you sign in to Blevo on a Blevo page and pick what to share there.

1. On gemini.google.com, open Settings›Connected Apps (on some accounts it’s Personal Intelligence›Connected Apps).
2. Under Custom apps, choose Add a custom app, paste https://mcp.blevo.app/mcp and click Next.
3. A Blevo page opens: sign in with the account you use in the app (or enter the code from You›Settings›AI›Connect AI assistants in the app), choose what to share, and approve.
4. In a chat, type @ and pick Blevo. Custom apps work in chat and with Gemini Spark.

Google limits custom apps to people aged 18 or over in the US, using a personal (not work or school) Google Account. Gemini asks you to confirm before an app changes anything.

1. Open Account settings›Connectors and click + Custom connector, then choose Remote.
2. Name it Blevo, paste https://mcp.blevo.app/mcp as the MCP server URL, choose OAuth for authentication and Streamable HTTP for transport. Tick the acknowledgement and click Add.
3. Click the Blevo card to connect. A Blevo page opens: sign in with the account you use in the app (or enter the code from You›Settings›AI›Connect AI assistants in the app), choose what to share, and approve.

On Enterprise, members can add their own connectors only if an admin allows it.

Mistral’s assistant Le Chat is now called Vibe.

1. Open Connectors from the sidebar, click + Add Connector and switch to the Custom MCP Connector tab.
2. Enter blevo as the connector name (no spaces), paste https://mcp.blevo.app/mcp as the server URL and click Connect.
3. A Blevo page opens: sign in with the account you use in the app (or enter the code from You›Settings›AI›Connect AI assistants in the app), choose what to share, and approve.

In team workspaces only admins can add connectors. We suggest leaving Blevo’s change functions on “ask” rather than “Always allow”.

1. Open your agent’s Tools page and choose Add a tool›New tool›Model Context Protocol.
2. Enter Blevo as the server name, a short description, and https://mcp.blevo.app/mcp as the server URL. For authentication choose OAuth 2.0 → Dynamic discovery, then Create.
3. Choose Create a new connection, sign in to Blevo, then Add to agent.

Blevo data is personal. Set the tool up so each person uses their own Blevo sign-in — never share one Blevo account across an agent’s users. Generative orchestration must be on for MCP tools.

Add Blevo as a remote (Streamable HTTP) MCP server. Each editor signs you in to Blevo in your browser the first time you use it.

{
  "mcpServers": {
    "blevo": {
      "url": "https://mcp.blevo.app/mcp"
    }
  }
}

{
  "servers": {
    "blevo": {
      "type": "http",
      "url": "https://mcp.blevo.app/mcp"
    }
  }
}

{
  "mcpServers": {
    "blevo": {
      "serverUrl": "https://mcp.blevo.app/mcp"
    }
  }
}

{
  "context_servers": {
    "blevo": {
      "url": "https://mcp.blevo.app/mcp"
    }
  }
}

codex mcp add blevo --url https://mcp.blevo.app/mcp
codex mcp login blevo

In Windsurf, open the file from Cascade’s MCP menu — its location depends on your version. In VS Code you can also run MCP: Add Server. Claude Code is on the Claude tab.

Model APIs call Blevo from their servers, so there’s no browser sign-in. Use a personal access token instead:

1. In the Blevo app, open You›Settings›AI›Connect AI assistants›Access tokens and create a token. Choose only the permissions you need.
2. Copy it (it starts with blevo_pat_) and keep it secret, like a password. Anyone with it can use what you allowed.
3. Send it as a bearer token to the connector URL. Revoke it in the same place in the app when you’re done — it stops working immediately.

Chat apps like ChatGPT, Claude and Gemini sign in with OAuth (the other tabs). Tokens are only for code you run yourself — never paste one into a chat.

curl https://mcp.blevo.app/mcp \
  -H "Authorization: Bearer blevo_pat_…" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

curl https://api.openai.com/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "gpt-6-astra",
    "input": "How did my week go?",
    "tools": [{
      "type": "mcp",
      "server_label": "blevo",
      "server_url": "https://mcp.blevo.app/mcp",
      "authorization": "blevo_pat_…",
      "require_approval": { "never": { "read_only": true } }
    }]
  }'

curl https://api.x.ai/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $XAI_API_KEY" \
  -d '{
    "model": "grok-4.7",
    "input": "How did my week go?",
    "tools": [{
      "type": "mcp",
      "server_label": "blevo",
      "server_url": "https://mcp.blevo.app/mcp",
      "authorization": "blevo_pat_…"
    }]
  }'

Swap in whichever model you use. OpenAI asks for approval before tool calls by default; the example skips it only for Blevo’s read-only tools. Any other MCP client works the same way: server URL https://mcp.blevo.app/mcp, header Authorization: Bearer blevo_pat_….