Action guide

Connect ChatGPT Or Claude To Logit

Add Logit as a connector in ChatGPT, Claude, or another MCP-capable assistant, approve exactly what it may read or add, and disconnect it any time from Settings.

• 6 min
• take action
• Add Logit as a connector and choose what it may do

You can let an assistant you already use, such as ChatGPT or Claude, work with your Logit: turn a conversation into a private journal event, look up what you planned this week, or add a task you committed to. You sign in to Logit from the assistant, choose what it may do, and it never gets a password or a key to copy.

The Logit connector is switched on per environment and reaches production behind a separate rollout decision. Until it is available to you, a pasted key from Self → Settings → Connected agents keeps working everywhere.

🎯 Use This When

• You want to log a conversation from ChatGPT or Claude into Logit.
• You want an assistant to see your plan, journal, or Circles while you talk.
• You want to check or end an assistant's access.

🔒 What Stays Private

• Your journal, plan, and Circles stay in Logit. The assistant receives only what a tool call returns for the areas you approved, and only while you are using it.
• Nothing connected can delete journal entries, actions, or discussions.
• Every connection is tied to your account alone. It never sees other people's private data or anything you did not approve.
• What the assistant's company keeps from your conversation is governed by that company's privacy policy, not by Logit.

✅ One Thing To Do Now

Add the connector and approve Read only first. Widen it later if you want the assistant to add or update on your behalf.

In short: you add Logit once, approve what it may see or add, and can end the connection whenever you like.

Add Logit to your assistant

ChatGPT

1. Open Settings → Connectors → Create (or, in the composer, **Developer mode → Add MCP server**).
2. Name it Logit and enter https://logit.social/mcp as the MCP server URL.
3. Choose OAuth as the authentication method. ChatGPT opens a Logit page.
4. Sign in to Logit if asked, read the permissions, and choose **Allow ChatGPT**.

Claude

1. In Claude, open Settings → Connectors → Add custom connector.
2. Name it Logit and enter https://logit.social/mcp.
3. Choose Connect. Claude opens a Logit page.
4. Sign in to Logit if asked, read the permissions, and choose Allow Claude.

Claude Code works the same way from a terminal:

claude mcp add --transport http logit https://logit.social/mcp

Then run /mcp, choose Logit, and choose Authenticate.

Other assistants

Any assistant that can add an MCP server with sign-in can connect the same way. Add https://logit.social/mcp and pick its sign-in or OAuth option.

What the consent page shows

The Logit page names the assistant and lists exactly what it asked for, area by area: read your journal, add to and update your goals, tasks, and habits, read your Circles, and so on. Two things are always true:

• No deleting. No assistant can delete journals, actions, or discussions through this connection.
• Read only is one switch. If the assistant asked to add or update, turn on Read only to let it read without changing anything. You can reconnect later to widen it.

Use it

Ask in plain words: "Log this conversation into Logit as a journal event", "What is on my plan today?", or "Add a task to call the dentist next Tuesday". The assistant proposes what it will create and, for ChatGPT and Claude, asks you before it changes anything. Anything an assistant adds is marked as drafted by that assistant, so you always know what you wrote yourself.

Check or end access

Open Self → Settings → Connected agents. Each connected assistant appears as one row showing when it connected, when it last used Logit, and what it may do in each area. Choose Disconnect to end its access immediately; the assistant cannot sign back in on its own and would have to be connected again from scratch. Disconnecting does not delete anything the assistant already saved into your journal or plan; those items are yours and live in Logit like anything else.

If something goes wrong

• The Logit page says the request could not be checked: choose **Try again. Logit only offers Allow** once it knows which app is asking and what it wants, so nothing is connected until then. If it keeps failing, sign in to Logit again and add the connector once more.
• The assistant says Logit is unavailable or asks you to reconnect: open Connected agents, disconnect the old row, and add the connector again.
• The assistant cannot see an area you expected: it may have been connected read only or without that area. Disconnect and connect again, approving what you want.
• You did not connect an assistant that appears in the list: disconnect it, then contact support through Self → Settings → Help & support.

➡️ Where To Go Next

• Meet Loggy, Your Journaling AI for the assistant that lives inside Logit and always confirms changes with you.
• Privacy And Data for what starts private and how sharing stays a deliberate choice.
• Deletion And Export for ending a connection, exporting, or leaving.

🌿 Why This Matters

Reflection only helps when it happens where you already are. If a conversation with an assistant is where a plan took shape, logging it there, privately and with your approval, keeps the record honest without adding another app to remember.

🧠 Related Concepts

• What Is Balance
• Memory And Meaning

🧭 Related Areas

• Reflect: the journal an assistant writes into.
• Act: the goals, tasks, and habits it can add.
• Self: where connected assistants are managed.

⭕ Related Life Rings

• Reflect Ring
• Act Ring

🛠️ Related Features

• Meet Loggy, Your Journaling AI
• The Daily Life Wheel Notebook
• How to Set Goals and Turn Them Into Action

📚 Continue Learning

• Your Daily Workflow to see where an assistant-logged event lands in your day.
• Privacy And Data to decide what, if anything, you later share from those entries.

❓ Frequently Asked Questions

Does the assistant get my Logit password? No. You sign in on a Logit page and it receives only a revocable credential tied to your account and the permissions you approved.

Can it delete things? No. Nothing connected this way can delete journal entries, actions, or discussions.

What does the assistant's company keep? Whatever it read or you asked it to add becomes part of your conversation with that assistant, under that company's privacy policy. Logit keeps only the connection details and content-free records of each call.

How do I stop it? Open Self → Settings → Connected agents and choose Disconnect. Access ends immediately.

Why does it say Logit is not available yet? The connector is switched on per environment and reaches production behind a separate rollout decision. A pasted key from Connected agents works in the meantime.