# LLMGRAPH for Gemini CLI

> Add the LLMGRAPH MCP server to Gemini CLI on Windows: the config file, the cmd /c npx form, and how to check it connected.

Add LLMGRAPH to Gemini CLI with one entry in `.gemini/settings.json`; if Gemini CLI cannot start `npx` on Windows, use `cmd /c npx -y llmgraph-mcp`. This guide follows Gemini CLI's own docs and is not yet tested by us.

## 1. Add the server to Gemini CLI

Add (from the vendor's docs, not yet tested by us): ~/.gemini/settings.json (user) or .gemini/settings.json (project)

```json
{
  "mcpServers": {
    "llmgraph": {
      "command": "npx",
      "args": [
        "-y",
        "llmgraph-mcp"
      ]
    }
  }
}
```

gemini mcp add exists (usage: gemini mcp add [options] <name> <commandOrUrl> [args...]); no command form is given here, so use the file entry below.

Gemini CLI's own docs do not mention a `cmd /c` wrapper on Windows. If it cannot start `npx`, use `cmd /c npx -y llmgraph-mcp`.

## 2. Check that it connected

Check (from the vendor's docs, not yet tested by us; your agent lists them as tools):
- Type /mcp inside Gemini CLI: it lists configured MCP servers and tools (list is the default action).

## 3. Sign in once

The first time your agent asks a question, LLMGRAPH answers with a sign-in link and a 6-character code. In a terminal you can run this instead:

```sh
npx -y llmgraph-mcp login
```

The link works for 10 minutes. Open the full link, including everything after the `#`, sign in with Google, compare the code on the page with the one your agent shows, and approve this device. Approving your first device starts your 7-day trial. The full steps are in [Set up with your agent](https://beta.llmgraph.in/docs/start.md).

## 4. Make sure Gemini CLI uses it

Adding a server doesn't make an agent use it.

Add a line to your agent's instructions file, `GEMINI.md`:

```text
Before changing code, ask the llmgraph MCP tools who_calls and what_breaks_if_changed; after editing, run check_my_changes; read each answer's label and its lines about what it cannot see.
```

We can't promise that an agent asks on its own. Every answer starts with `llmgraph · `: look for it where Gemini CLI shows what it asked.

## 5. The five questions

| Question | What it answers |
| --- | --- |
| [who_calls](https://beta.llmgraph.in/docs/questions/who_calls.md) | Every place that calls this function, plus calls we can't see. |
| [what_breaks_if_changed](https://beta.llmgraph.in/docs/questions/what_breaks_if_changed.md) | Everything that depends on this, directly or indirectly, and what we can't see. |
| [check_my_changes](https://beta.llmgraph.in/docs/questions/check_my_changes.md) | After editing: callers you didn't update, and anything we can't see. |
| [what_is_unknown](https://beta.llmgraph.in/docs/questions/what_is_unknown.md) | Where our view is blind (dynamic calls, unknown types, unreadable files) and why. |
| [explain](https://beta.llmgraph.in/docs/questions/explain.md) | Why we said this: the evidence behind one answer line. |

## Other agents

- [LLMGRAPH for Claude Code](https://beta.llmgraph.in/docs/agents/claude-code.md): MCP setup on Windows.
- [LLMGRAPH for Cursor](https://beta.llmgraph.in/docs/agents/cursor.md): MCP setup on Windows.
- [LLMGRAPH for Codex](https://beta.llmgraph.in/docs/agents/codex.md): MCP setup on Windows.
- [LLMGRAPH for Antigravity](https://beta.llmgraph.in/docs/agents/antigravity.md): MCP setup on Windows.
- [LLMGRAPH for OpenCode](https://beta.llmgraph.in/docs/agents/opencode.md): MCP setup on Windows.
- [LLMGRAPH for Windsurf](https://beta.llmgraph.in/docs/agents/windsurf.md): MCP setup on Windows.

- [Set up with your agent](https://beta.llmgraph.in/docs/start.md): Every agent on one page, plus the sign-in steps and what to do when something goes wrong.
- [The MCP server](https://beta.llmgraph.in/docs/mcp-server.md): What the server answers and what it needs.
- [Questions and answers](https://beta.llmgraph.in/docs/faq.md): Short answers to common questions.