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

# Quickstart: Add Lava to Your AI Chat

> Install the Lava MCP in your AI chat, connect a service, and make your first call in under 5 minutes.

The Lava MCP gives your AI chat access to a catalog of connected services with auth and billing handled automatically. This quickstart walks you from zero to a successful call in under 5 minutes.

By the end, you'll have Lava installed in your chat, one service connected, and one call returning real data.

## Step 1: Install Lava in Your AI Chat

Click [Add Lava to claude.ai](https://claude.ai/settings/connectors?modal=add-custom-connector\&connectorName=Lava\&connectorUrl=https://www.lava.so/mcp). claude.ai opens its connector dialog with Lava as the connector name and `https://www.lava.so/mcp` as the URL, and asks you to confirm. Approve it.

**Expected result.** Lava appears in your claude.ai connectors list. The first time you use it, claude.ai prompts you to sign in (covered in Step 2).

**Other surfaces.** Using Claude Desktop, Claude Code, or installing for a Team or Enterprise workspace? See the [install guide](/mcp/install) before continuing.

## Step 2: Sign In to Lava

The first time your AI chat uses a Lava tool, Lava redirects you to a browser to sign in. Complete the sign-in flow and return to your chat. In Claude Code, run `/mcp` and follow the prompts to authenticate `lava`. For details on what to expect, see [Sign In and Confirm](/mcp/install#sign-in-and-confirm) in the install guide.

**Expected result.** Your AI chat confirms the connection. Lava is now ready to handle tool calls on your behalf.

## Step 3: Connect Your First Service

Connect web search. In your chat, type:

```text theme={null}
Connect web search to Lava.
```

Your chat hands off to Lava, which replies with a sign-in link:

```json theme={null}
// Lava sends back:
{
  "url": "https://www.lava.so/auth/connect/serper?session=...",
  "expires_in": 600
}
```

Open the URL. Lava walks you through whichever sign-in the search provider needs, then returns you to your chat.

**Expected result.** Ask your chat what you have connected, and the search provider now appears:

```text theme={null}
What services do I have connected?
```

```json theme={null}
// Lava sends back:
{
  "data": [
    { "service": "serper", "connected_at": "2024-01-01T00:00:00Z" }
  ]
}
```

## Step 4: Make Your First Call

Now have Lava use the connected service. Type:

```text theme={null}
Search the web for tips for visiting Lisbon for the first time and summarize the top three results.
```

Your chat picks the right Lava tool — typically `call_api` against a search provider — and Lava handles auth and billing.

**Expected result.** Your chat returns a summary along these lines (actual content varies — Lava searches the web at query time):

```text theme={null}
Top three first-time tips for visiting Lisbon:

1. Pack comfortable shoes. Lisbon's old neighborhoods are built on steep
   hills and laid with cobblestone — sneakers or sturdy walking shoes
   beat fashion choices.

2. Use the historic Tram 28 early or late. It runs through the most
   scenic neighborhoods, but fills up by mid-morning. Going at 8am or
   after 7pm gets you a seat and better photos.

3. Try a pastel de nata at Pastéis de Belém. The custard tarts here are
   the original recipe; expect a line, but it moves quickly.
```

If your chat tells you Lava can't run that yet because of an insufficient balance, continue to Step 5.

## Step 5: Add Funds to Unlock Paid Providers

Free providers work without funds. Most paid providers require a wallet balance. Ask your chat:

```text theme={null}
Add $25 to my Lava wallet.
```

Lava returns either a list of saved payment methods (if you have one) or a browser checkout URL. Follow the flow to add funds.

**Expected result.** Asking your chat:

```text theme={null}
What's my Lava balance?
```

returns the new total.

For deeper coverage — including the dashboard funding flow and saved payment methods — see [How to Fund Your Lava Account](/mcp/funds).

## What You've Done

You installed Lava in your AI chat, signed in, connected a service, and made your first call. Your chat can now reach any service in Lava's catalog through the same set of tools.

## Next Steps

* [How to Install the Lava MCP](/mcp/install) — Claude Desktop, Claude Code, and Team or Enterprise installs.
* [How to Connect Services through Lava MCP](/mcp/connect-services) — add Gmail, Slack, Notion, and more.
* [How to Fund Your Lava Account](/mcp/funds) — funding, balance, usage history.
* [Lava MCP Tool Reference](/mcp/tool-reference) — every tool the MCP exposes, with parameters and examples.