> ## Documentation Index
> Fetch the complete documentation index at: https://docs.withgauge.com/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> For API and SDK launches, integration testing, and ongoing coding-agent checks, start with Gauge Agents:
> - [Gauge Agents overview](https://docs.withgauge.com/agents/index.md)
> - [Gauge Agents quickstart](https://docs.withgauge.com/agents/quickstart.md)
> - [Create and inspect an eval](https://docs.withgauge.com/agents/guides/create-an-eval-and-run-an-experiment.md)
> - [Agent Experience: diagnose integration failures](https://docs.withgauge.com/agents/concepts/agent-experience.md)
> - [Agent Preference: measure product selection](https://docs.withgauge.com/agents/concepts/agent-preference.md)
> - [GitHub PR checks](https://docs.withgauge.com/agents/guides/github-pr-checks.md)
> For AI search visibility and AEO, start with [Gauge Chat](https://docs.withgauge.com/introduction/welcome.md).

# Connect Gauge to your tools

> Use Gauge Chat and Gauge Agents from your AI assistant, terminal, coding agent, or applications with MCP, the CLI, or the API.

Use Gauge from your AI assistant, terminal, or applications. All three options work with Gauge Chat and Gauge Agents.

| Option | Best for |
| - | - |
| [MCP](#mcp) | Asking questions and taking actions from an AI assistant |
| [CLI](#cli) | Terminal work, scripts, CI, and coding agents |
| [API](#api) | Custom apps and automations |

## MCP

**Use Gauge from your AI assistant.** Connect Claude, Cursor, ChatGPT, Codex, or another MCP client to one Gauge server, then ask for what you need in plain language.

* **Gauge Chat:** analyze visibility and citations, manage topics and prompts, create content, and work the Action Center.
* **Gauge Agents:** review evals and agent sessions, read run analysis, and launch saved evals.

For example, ask your assistant:

> How did our visibility in ChatGPT change over the last 30 days, and which competitors gained? Then show why the latest run of our onboarding eval failed.

Most clients connect with OAuth: you approve access in the browser and choose an organization.

> [card] [Set up Gauge MCP](https://docs.withgauge.com/gauge-mcp)
>
> Connect your client with OAuth or an API key, and see the tool list.

## CLI

**Use Gauge from your terminal.** Run `gauge` commands by hand, in scripts, or in CI jobs. Every command supports `--help`, and `-o json` returns output for scripts.

* **Gauge Chat:** `gauge chat` reads categories, brands, answers, and citations, and manages topics, prompts, and tags.
* **Gauge Agents:** run evals locally in Docker or on hosted Gauge, analyze sessions, and compare results.

```bash theme={null}
gauge chat citations list --category <category> --since 2026-09-01 -o json
```

### Use the CLI from a coding agent

The CLI also works well with coding agents such as Claude Code and Codex. Install the `gauge-agents` skill so your agent knows how to use Gauge in future sessions:

```bash theme={null}
gauge install-skill
```

Before its first `gauge` command, your agent runs `gauge instructions`. This command prints guidance written for coding agents, and the guidance ships inside the CLI, so it always matches your installed version. The `hillclimb` topic covers the core improvement loop: run a pilot, find where agents get stuck, fix your docs or Skills, and rerun the test.

```bash theme={null}
gauge instructions             # overview and the improvement loop
gauge instructions hillclimb   # full guidance for the loop
gauge instructions --list      # all topics
```

> [card] [Install and sign in to the CLI](https://docs.withgauge.com/cli)
>
> Install the CLI, sign in or use an API key, and choose an organization.

## API

**Build Gauge into your workflows.** Use the REST API to connect custom apps, internal dashboards, and automations.

* **Gauge Chat API:** manage tracked prompts, topics, and tags, and query AI visibility reports.
* **Gauge Agents API:** manage evals, launch runs, inspect evidence, and query analytics.

For example, list the tracked prompts in a Gauge Chat category:

```bash theme={null}
curl "https://app.withgauge.com/api/v1/orgs/acme/categories/crm-software/prompts?limit=50" \
  --header "Authorization: Bearer $GAUGE_API_TOKEN"
```

One organization API key works with both APIs, the CLI, and MCP.

> [card] [Gauge Chat API](https://docs.withgauge.com/chat/api/authentication.md)
>
> Authenticate, then [make your first requests](https://docs.withgauge.com/chat/api/use-the-api.md) or read the [API reference](https://docs.withgauge.com/chat/api-reference/overview.md).

> [card] [Gauge Agents API](https://docs.withgauge.com/agents/guides/api-authentication.md)
>
> Authenticate, then [make your first requests](https://docs.withgauge.com/agents/guides/use-the-api.md) or read the [API reference](https://docs.withgauge.com/agents/api-reference/overview.md).

## One account for both products

Gauge Chat and Gauge Agents share one account, organization, and team. One MCP connection and one CLI sign-in cover both products, and the organization you select applies to both. See [Using Gauge Chat and Gauge Agents](https://docs.withgauge.com/introduction/chat-and-agents.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.