dashly

October 16, 2026 · MCP, Using Claude

What Is an MCP Dashboard? A Live Wall for Your AI

An MCP dashboard is a dashboard your AI assistant can read and write through the Model Context Protocol, the open standard for connecting assistants to apps. You talk to Claude or ChatGPT, it updates the dashboard, and you see the result as charts and tiles. Dashly is an MCP dashboard for the things you track in your life and work, from calories and habits to tasks and sales.

MCP in one paragraph

MCP gives an AI assistant a standard way to use another app. The app describes what it can do, such as “log an entry” or “summarize a topic”, and the assistant calls those actions when a conversation needs them. Claude, ChatGPT and Claude Code all support it, which means one MCP app works in all three. For you, it shows up as a connector you add once in your assistant’s settings.

What a dashboard over MCP does

Most MCP apps let the assistant fetch something: a file, a calendar, a database row. A dashboard flips the direction. The assistant writes to it as you talk, and the dashboard turns what it wrote into something you can see without asking.

Say “just had 16 ounces of water” and Claude logs it to your water topic. The water tile on your Dashly wall moves from 24 to 40 ounces of your 64. Say “finished the panel swap at the Miller house” and the task closes on your job sites tile. You never opened Dashly, and it is up to date.

The assistant can read it back too. Ask “how did I sleep this week compared with last?” and it summarizes the topic from Dashly instead of guessing from the chat.

Why this is a better shape for tracking

Before MCP, tracking with an assistant meant copying its answers into a spreadsheet or a notes app by hand, and most people gave up on that within a few weeks.

An MCP dashboard removes the glue. The assistant handles the messy part: understanding “a large latte and a banana”, working out the numbers, picking the right topic. The dashboard handles the part assistants are bad at: keeping exact values with dates, adding them up and drawing them over months.

It also means you are not locked to one assistant. Dashly lives at one address, so Claude on your laptop, ChatGPT on your phone and Claude Code in your terminal all write to the same wall.

What Dashly gives you

Every topic you track has a type, a unit and an optional goal or limit. Dashly draws each one as a live tile that fits it: a glass filling for water, a streak for habits, a list that keeps moving for tasks. A Today tile scores the whole day. Each topic has its own page with its full history, and you can put a dashboard on a TV or a tablet by pairing it with a code.

The features on the home page show these tiles working with sample data, the same components the app uses.

How to connect it

  1. In Claude, open Settings, then Connectors, and add a custom connector with the address https://dashly.si/mcp. The Claude guide has each step, and the Claude Code guide has the one-line command.
  2. Sign in to Dashly when Claude asks and approve the connection.
  3. Tell Claude what to track.

What to track first

Start with something you already mention to your assistant. For most people that is food, water, workouts or a to-do list. One or two topics are enough to see how it works, and you can add more by asking whenever you like.

Questions

  • What is MCP?

    The Model Context Protocol is an open standard for connecting AI assistants to outside tools and data. An app that speaks MCP can be added to Claude, ChatGPT and other assistants as a connector.

  • Which assistants can use an MCP dashboard?

    Claude, ChatGPT and Claude Code, and any other assistant that supports MCP connectors. Dashly works with all of them at the same address.

  • Do I need to be a developer to use Dashly over MCP?

    No. Adding Dashly is a few clicks in your assistant's settings and a sign-in. After that you only talk to your assistant.

  • What can my assistant do in Dashly?

    Create topics with goals, log and correct entries, read summaries and build dashboards, within the permissions you approve when you connect.