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

# Codex Docs MCP

> Add the Codex documentation MCP server to your AI coding tools for instant access to API docs.

The Codex documentation is available as a [Model Context Protocol](https://modelcontextprotocol.io/) (MCP) server. Adding it to your AI coding tool lets it search and reference the Codex API docs directly, so you can get accurate answers without leaving your editor.

```
https://docs.codex.io/mcp
```

## What it does

Once connected, your AI tool can search the full Codex API documentation — including queries, subscriptions, types, and guides. It will automatically reference the docs when you ask questions about the Codex API or need help writing queries.

<Info>
  This MCP server is for searching the documentation and helping your AI write queries. It does not interface with the Codex API directly — you'll still need to execute queries using your API key via the [GraphQL endpoint](/learn-graphql) or the [SDK](/sdk).
</Info>

## Docs MCP vs Codex Skills

The [Codex Skills](/agents/codex-skills) gives your AI agent preloaded knowledge of the API so it can interface with the API directly. The Docs MCP lets it search the documentation, which can allow it to write GraphQL queries.

|                    | Docs MCP                                     | Codex Skills                                                              |
| ------------------ | -------------------------------------------- | ------------------------------------------------------------------------- |
| **How it works**   | Searches docs on demand via MCP protocol     | Preloaded API knowledge — operations, auth, templates                     |
| **Best for**       | Looking up specific fields, types, or guides | Interacting with the Codex API directly                                   |
| **Needs network?** | Yes — queries the MCP server                 | Works offline to write queries (still needs a connection to call the API) |

Use both together for the best experience: the skill handles query generation and the MCP fills in details when needed.

## Setup

<Tabs>
  <Tab title="Cursor">
    Open MCP settings with `Cmd+Shift+P` (Mac) or `Ctrl+Shift+P` (Windows/Linux) and select **Cursor Settings: Open MCP Settings**. Add the following:

    ```json theme={null}
    {
      "mcpServers": {
        "codex-docs": {
          "url": "https://docs.codex.io/mcp"
        }
      }
    }
    ```
  </Tab>

  <Tab title="VS Code">
    Create a `.vscode/mcp.json` file in your project root:

    ```json theme={null}
    {
      "servers": {
        "codex-docs": {
          "type": "http",
          "url": "https://docs.codex.io/mcp"
        }
      }
    }
    ```
  </Tab>

  <Tab title="Windsurf">
    Open the MCP configuration file at `~/.codeium/windsurf/mcp_config.json` and add the following:

    ```json theme={null}
    {
      "mcpServers": {
        "codex-docs": {
          "serverUrl": "https://docs.codex.io/mcp"
        }
      }
    }
    ```
  </Tab>

  <Tab title="Claude Desktop">
    Go to **Settings > Connectors**, select **Add custom connector**, then enter:

    * **Name:** Codex Docs
    * **URL:** `https://docs.codex.io/mcp`
  </Tab>

  <Tab title="Claude Code">
    Run the following command:

    ```bash theme={null}
    claude mcp add --transport http codex-docs https://docs.codex.io/mcp
    ```
  </Tab>
</Tabs>

## Example prompts

Once the MCP is connected, you can ask your AI assistant to build features using the Codex API. Here are some examples to get started.

### Get a token price

> "Use the Codex API to fetch the current USD price of WETH on Ethereum."

Your assistant will look up the `getTokenPrices` query and write something like:

```graphql theme={null}
query {
  getTokenPrices(
    inputs: [
      {
        address: "0xc02aaa39b223fe8d0a0e5c4f27ead9083c756cc2"
        networkId: 1
      }
    ]
  ) {
    address
    priceUsd
  }
}
```

### Build a trending tokens feed

> "Show me how to get the top 10 trending tokens on Base sorted by 24h volume."

```graphql theme={null}
query {
  filterTokens(
    filters: { network: [8453] }
    rankings: { attribute: volume24, direction: DESC }
    limit: 10
  ) {
    results {
      token {
        name
        symbol
        address
      }
      volume24
      priceUSD
      change24
    }
  }
}
```

### Render a price chart

> "Fetch 1-hour OHLCV candles for the last 24 hours for this token on Solana: `So11111111111111111111111111111111111111112`."

```graphql theme={null}
query {
  getBars(
    symbol: "So11111111111111111111111111111111111111112:1399811149"
    from: 1719792000
    to: 1719878400
    resolution: "60"
  ) {
    o
    h
    l
    c
    volume
    t
  }
}
```

### Stream live trades

> "Set up a WebSocket subscription to stream real-time swap events for a token on Ethereum."

```graphql theme={null}
subscription {
  onEventsCreated(
    address: "0xc02aaa39b223fe8d0a0e5c4f27ead9083c756cc2"
    networkId: 1
  ) {
    address
    networkId
    events {
      eventDisplayType
      timestamp
      maker
      token0SwapValueUsd
      token1SwapValueUsd
      transactionHash
    }
  }
}
```

### Track wallet holdings

> "Get all token balances for a wallet on Base."

```graphql theme={null}
query {
  balances(
    input: {
      walletAddress: "0xd8dA6BF26964aF9D7eEd9e03E53415D37aA96045"
      networks: [8453]
    }
  ) {
    items {
      tokenAddress
      balance
      shiftedBalance
      token {
        name
        symbol
      }
    }
  }
}
```

### Discover new launches

> "Find tokens launched in the last hour on Solana with at least \$10k liquidity."

```graphql theme={null}
query {
  filterTokens(
    filters: {
      network: [1399811149]
      liquidity: { gte: 10000 }
      createdAt: { gte: 1719874800 }
    }
    rankings: { attribute: createdAt, direction: DESC }
    limit: 20
  ) {
    results {
      token {
        name
        symbol
        address
      }
      liquidity
      priceUSD
      createdAt
    }
  }
}
```

<Tip>
  You don't need to memorize any of this — just describe what you want and your AI assistant will look up the right query from the docs. The examples above are just to show what's possible.
</Tip>