> For the complete documentation index, see [llms.txt](https://docs.scribeware.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.scribeware.com/into-the-weeds/ai-assistant-mcp/connecting-your-ai-tool.md).

# Connecting Your AI Tool

Switch the AI Assistant on, then paste one snippet into Claude Desktop, Claude Code, Gemini CLI or whatever else you use.

## Step 1 — Switch It On

1. Open **Settings → AI**.
2. Turn on **Enable AI Assistant (MCP)**.

The panel expands to show a status chip and a set of setup snippets. The chip should read **Running on port 8765** (or a nearby number). If it says **Stopped — log in to start**, log in to ScribeWare and it'll start on its own.

That switch is remembered on this computer, not on your account — so it survives logging out, and turning it on here doesn't turn it on for the rest of your team. Each person switches it on for themselves.

{% hint style="info" %}
Port 8765 is the usual one, but if something else on your machine has already claimed it, ScribeWare quietly takes the next free port. Always copy the snippet from the panel rather than typing the address from memory — it always shows the port actually in use.
{% endhint %}

## Step 2 — Point Your AI Tool at It

Each snippet in the panel has a **copy** button in its top-right corner. Copy the one for your tool, paste it where indicated, and restart the tool.

### Claude Desktop

Claude Desktop connects through a small bridge script that ships inside ScribeWare, so its snippet is a little longer than the others. It looks like this, with real paths filled in for your machine:

```json
{
  "mcpServers": {
    "scribeware": {
      "command": "/path/to/ScribeWare",
      "args": ["/path/to/mcp-proxy.js"],
      "env": { "ELECTRON_RUN_AS_NODE": "1" }
    }
  }
}
```

Merge it into your Claude Desktop config file:

* **macOS** — `~/Library/Application Support/Claude/claude_desktop_config.json`
* **Windows** — `%APPDATA%\Claude\claude_desktop_config.json`

Then **quit Claude Desktop completely and reopen it**. Claude only reads that file at startup.

{% hint style="info" %}
"Merge" matters if you already use other MCP servers. Add the `scribeware` entry inside your existing `mcpServers` block rather than replacing the whole file, or you'll disconnect everything else.
{% endhint %}

### Claude Code

```json
{
  "mcpServers": {
    "scribeware": {
      "type": "http",
      "url": "http://127.0.0.1:8765/mcp"
    }
  }
}
```

Add it to `~/.claude/settings.json` (macOS) or `%USERPROFILE%\.claude\settings.json` (Windows).

### Gemini CLI

```json
{
  "mcpServers": {
    "scribeware": {
      "httpUrl": "http://127.0.0.1:8765/mcp"
    }
  }
}
```

Add it to `~/.gemini/settings.json` (macOS) or `%USERPROFILE%\.gemini\settings.json` (Windows).

### Something Else

Most other AI tools connect over plain HTTP. Give it the URL shown in the panel — `http://127.0.0.1:8765/mcp` — wherever it asks for an MCP server or SSE endpoint.

If you can't work out where that goes in your particular tool, copy the Claude Code snippet, paste it into the AI itself and ask it how to configure an MCP server for the tool you're using. This is the kind of question they're good at.

## Step 3 — Check It Worked

Open a report in ScribeWare, then ask your AI something that only ScribeWare can answer:

> What chapters are in the report I have open?

A correct answer means everything's wired up. If it says it can't see any ScribeWare tools, or asks you for a report ID, see [Troubleshooting](/into-the-weeds/ai-assistant-mcp/troubleshooting-the-ai-assistant.md).

## Keeping It Connected

* **ScribeWare has to be running.** The AI reaches your data through the desktop app. If it's closed, the tools simply aren't there.
* **You have to be logged in.** Logging out shuts the connection down, and logging back in brings it up again.
* **You don't need to reconnect each day.** Set it up once; it comes back with the app.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.scribeware.com/into-the-weeds/ai-assistant-mcp/connecting-your-ai-tool.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.