---
title: PromptEasy MCP Server
description: Integrate PromptEasy directly into your AI workflows and agents using the Model Context Protocol.
---

PromptEasy acts as a Model Context Protocol (MCP) Server, allowing AI assistants to interact with the PromptEasy API to manage prompts, vaults, and other resources.

## Authentication

To authenticate with the PromptEasy MCP Server, you must provide a Bearer Token (either personal access token or service account token). You can generate these tokens after login.

### Token Types

- **Personal Access Tokens**: All users can generate personal access tokens for their own AI tools to connect to the MCP server. These tokens are subject to the standard API rate limits.

- **Service Accounts**: For our **Premium** and **Enterprise** customers, we offer Service Accounts (bot users). These accounts are designed for automated, high-volume access and benefit from a rate limit that is **10x higher** than regular user accounts.

### Example Configuration

Here is an example of how to configure your MCP client to connect to the PromptEasy server:

```json
{
  "mcpServers": {
    "prompteasy": {
      "httpUrl": "https://api.prompteasy.eu/v1/mcp/server",
      "headers": {
        "Authorization": "Bearer <YOUR_PERSONAL_ACCESS_TOKEN_OR_SERVICE_ACCOUNT_TOKEN>"
      }
    }
  }
}
```

## Capabilities

The MCP Server exposes the following tools:

### `list_vaults`

- **Description**: List all vaults the current user has access to.
- **Input**: None.
- **Output**: A JSON list of vault objects, each with an `id` and `name`.

### `search_prompts`

- **Description**: Search for prompts within accessible vaults.
- **Input**:
    - `query` (string, required): The search query.
    - `vault_id` (integer, optional): The ID of a specific vault to search within.
- **Output**: A JSON list of prompt objects, each containing `id`, `title`, `keywords`, and a nested `vault` object (`id`, `name`).

### `get_prompt`

- **Description**: Get details of a specific prompt by its ID.
- **Input**:
    - `id` (string, required): The UUID of the prompt to retrieve.
- **Output**: A JSON object containing `id`, `title`, `prompt`, `llm_model`, `llm_system_message`, `llm_settings`, `parameters`, `created_at`, and a nested `vault` object (`id`, `name`).