npx mcpbar@latest edit -c claude

Note

This project is still under development and not all tools are available. We are working on adding more tools and improving the server. We are open to contributions! See CONTRIBUTING.md for more details.

Overview

This server provides a bridge between AI agents and the Akash Network, allowing AI models to deploy applications, create leases, manage deployments, and interact with other Akash services directly through typed tools. It implements the Model Context Protocol, making it compatible with various AI platforms and tools.

Features

• Wallet and Client Management: Handles Akash wallet authentication and client initialization
• Certificate Management: Manages Akash certificates
• Tools for Akash Interaction:
  • Account address retrieval
  • Deployment creation and management
  • SDL (Stack Definition Language) operations
  • Bid management
  • Lease creation
  • Manifest deployment

Prerequisites

• Node.js (v18 or later)
• An Akash wallet (mnemonic required for interaction with Akash Network)
• Basic knowledge of Akash Network

Installation

1. Clone the repository:

   git clone https://github.com/akash-network/mcp.git
   cd mcp
   

2. Install dependencies:

   npm install
   

3. Build the project:

   npm run build
   

Usage with Claude Desktop or Cursor

Add the following to your claude_desktop_config.json or .cursor/mcp.json. See here for more details.

Make sure to replace <path to dist/index.js> with the path to the dist/index.js file you built.

{
  "mcpServers": {
    "Akash": {
      "command": "node",
      "args": ["<path to dist/index.js>"],
      "env": {
        "AKASH_MNEMONIC": "<your mnemonic here>",
        "AKASH_RPC_URL": "https://rpc.akashnet.net:443" // optional, defaults to https://rpc.akashnet.net:443
      }
    }
  }
}

Usage

Development Mode

Run the server with the MCP inspector for development:

npm run dev

For watch mode during development:

npm run dev:watch

Using the Server with AI Models

The server exposes a standard MCP interface that can be used by AI models to interact with the Akash Network. It uses the stdio transport by default, making it compatible with most MCP clients.

Available Tools

The server provides the following tools for AI agents:

• GetAccountAddrTool: Retrieve your Akash account address
• GetBalancesTool: Get the AKT (uakt) and other balances for a given Akash account address
• GetBidsTool: Get bids for deployments
• CreateDeploymentTool: Create a new deployment on Akash Network
• AddFundsTool: Deposit additional AKT (uakt) into a deployment escrow account
• GetSDLsTool: Get a list of available SDLs (from awesome-akash repository)
• GetSDLTool: Get a specific SDL by name
• SendManifestTool: Send a manifest to a provider
• CreateLeaseTool: Create a lease with a provider
• GetServicesTool: Get information about active services
• UpdateDeploymentTool: Update a deployment on Akash Network

GetBalancesTool

Description:

Get the AKT (uakt) and other balances for a given Akash account address.

Input Schema:

{
  "address": "akash1..." // Akash account address (string, required)
}

Example Usage:

Request:

{
  "address": "akash1xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}

Response:

[
  {
    "denom": "uakt",
    "amount": "123456789"
  }
  // ...other tokens if present
]

AddFundsTool

Description:

Deposit additional AKT (uakt) into a deployment escrow account.

Input Schema:

{
  "address": "akash1...", // Akash account address (string, required)
  "dseq": 123456,          // Deployment sequence number (integer, required)
  "amount": "1000000"     // Amount to add in uakt (string, required)
}

Example Usage:

Request:

{
  "address": "akash1xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "dseq": 123456,
  "amount": "1000000"
}

Response:

"...transaction raw log or error message..."

Development

Linting and Formatting

# Run linter
npm run lint

# Fix linting issues
npm run lint:fix

# Format code
npm run format

License

See LICENSE

Acknowledgements

• Akash Network
• Model Context Protocol
• awesome-akash (included as a submodule)