# Installation

## Installation

There are two ways to install the ChainGPT Developer Kit for Claude Code: via the plugin installer (recommended) or via manual git clone.

### Plugin install (recommended)

Run this command inside Claude Code:

```
/plugin install ChainGPT-org/chaingpt-claude-skill
```

This automatically sets up the skill files, configures the MCP server via `.mcp.json`, and makes all 5 skills and 12 MCP tools available immediately. The only additional step is setting your API key (see API Key Setup).

### Manual install (git clone)

Clone the repository into your project's `.claude/skills/` directory:

```bash
git clone https://github.com/ChainGPT-org/chaingpt-claude-skill .claude/skills/chaingpt
```

For a user-level install that applies to all projects, clone to your home directory instead:

```bash
git clone https://github.com/ChainGPT-org/chaingpt-claude-skill ~/.claude/skills/chaingpt
```

#### Configure the MCP server (manual installs only)

If you installed via git clone, you need to build the MCP server and add it to your settings:

```bash
cd .claude/skills/chaingpt/mcp-server
npm install && npm run build
```

Then add the following to your `.claude/settings.json`:

```json
{
  "mcpServers": {
    "chaingpt": {
      "command": "node",
      "args": [".claude/skills/chaingpt/mcp-server/dist/index.js"],
      "env": { "CHAINGPT_API_KEY": "your-key-here" }
    }
  }
}
```

### What gets installed

After installation, the following components are available:

| Component         | Description                                                                           |
| ----------------- | ------------------------------------------------------------------------------------- |
| 5 skills          | Main API reference, playground, debug assistant, hackathon scaffolder, update manager |
| 12 MCP tools      | Direct API access from Claude conversations                                           |
| 16 reference docs | Complete API documentation for all 7 products                                         |
| 10 templates      | Production-ready project scaffolds                                                    |
| 45+ patterns      | Audited Solidity smart contract patterns                                              |
| 8 examples        | Working code in JavaScript and Python                                                 |
| Mock server       | Zero-credit local testing server                                                      |

### Verifying the installation

Open Claude Code and ask it anything about ChainGPT:

```
"What ChainGPT products are available and how much do they cost?"
```

If the skill is loaded correctly, Claude will respond with detailed product information including exact credit costs, SDK package names, and endpoint URLs — not generic information from its training data.

### Updating

To check for and apply updates, use the update skill inside Claude Code:

```
/chaingpt-update
```

This fetches the latest changes from GitHub and shows a changelog before applying. For manual installs, you can also run `git pull` in the skill directory.

### Prerequisites

<table><thead><tr><th width="191.0546875">Requirement</th><th>Details</th></tr></thead><tbody><tr><td>Claude Code</td><td><a href="https://code.claude.com">code.claude.com</a></td></tr><tr><td>ChainGPT API key</td><td><a href="https://app.chaingpt.org">app.chaingpt.org</a> — connect a wallet to sign up</td></tr><tr><td>API credits</td><td><a href="https://app.chaingpt.org/addcredits">Buy credits</a> — 1,000 credits = $10</td></tr><tr><td>Node.js</td><td>Required for the MCP server (LTS version recommended)</td></tr></tbody></table>


---

# Agent Instructions: 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:

```
GET https://docs.chaingpt.org/dev-docs-b2b-saas-api-and-sdk/chaingpt-claude-skill-and-plugin/getting-started/installation.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.