mcpField
L
leandrogavidia-vechain-mcp-server

leandrogavidia-vechain-mcp-server

Remote
github.com/leandrogavidiaFinance5Updated 56w ago
Repository

About leandrogavidia-vechain-mcp-server

leandrogavidia-vechain-mcp-server connects MCP clients to VeChain documentation, Thor API data, and wallet operations. It is for working with VeChain accounts, transactions, fees, tokens, and signatures through a set of query and signing tools.

Key features

  1. 01Searches official VeChain documentation
  2. 02Retrieves account, transaction, and block data
  3. 03Suggests priority fees
  4. 04Creates wallets and signs messages or transactions
  5. 05Queries balances, token details, and allowances
  6. 06Sends tokens and manages approvals

Use cases

  1. 01Look up VeChain API documentation
  2. 02Inspect accounts, transactions, or blocks
  3. 03Check token balances and allowances
  4. 04Sign data or send tokens on VeChain

MCP support

Exposes documentation search, Thor REST API queries, wallet and signature operations, and token tools to MCP clients. Supports stdio or Streamable HTTP operation.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://server.smithery.ai/@leandrogavidia/vechain-mcp-server/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

All endpoints
Streamable HTTPhttps://server.smithery.ai/@leandrogavidia/vechain-mcp-server/mcpAuth required

Tools

Details

From the project README
VeChain logo

VeChain MCP Server

VeChain MCP Server is an MCP server specifically designed for the VeChain network. It provides advanced functionalities such as querying the official VeChain documentation, executing HTTP requests to the Thor REST API in both Mainnet and Testnet environments, and managing cryptographic signatures through an integrated wallet.

Integration

Register the server in your MCP-aware host configuration.

{
  "mcpServers": {
    "vechainMcp": {
      "url": "https://server.smithery.ai/@leandrogavidia/vechain-mcp-server/mcp",
      "type": "streamable-http"
    }
  }
}

Tools

  • Vechain Docs
    • Docs

      • search_documentation: Search VeChain Documentation.
  • Thorest API
    • Accounts

      • get_account: Retrieve account details.
    • Transactions

      • get_transaction: Retrieve a transaction by ID.
    • Blocks

      • get_block: Get a VeChain block.
    • Fees

      • get_priority_fee: Suggest a priority fee.
  • Wallet & signatures
    • Wallet

      • create_wallet: Create a VeChain wallet (mnemonic + keys).
    • Signatures

      • sign_certificate: Create and sign a canonical certificate.

      • sign_raw_transaction: Sign raw transaction.

  • Goat SDK (VeChain Tools)
    • get_address: Get the address of the wallet

    • get_chain: Get the chain of the wallet

    • sign_message: Sign a message with the wallet

    • get_balance: Get the balance of the wallet for native currency or a specific ERC20 token.

    • get_token_info_by_ticker: Get information about a configured token (like contract address and decimals) by its ticker symbol.

    • convert_to_base_units: Convert a token amount from human-readable units to its smallest unit (e.g., wei).

    • convert_from_base_units: Convert a token amount from its smallest unit (e.g., wei) to human-readable units.

    • sign_typed_data_evm: Sign an EIP-712 typed data structure (EVM)

    • get_token_allowance_evm: Get the allowance of an ERC20 token for a spender (returns amount in base units)

    • send_token: Send native currency or an ERC20 token to a recipient, in base units.

    • approve_token_evm: Approve an amount (specified in base units) of an ERC20 token for a spender

    • revoke_token_approval_evm: Revoke approval for an ERC20 token from a spender (sets allowance to 0)


.env Config

  • AGENT_SECRET_KEY: CSecret key in string format that allows your MCP server to use the signatures tools.

  • ENVIRONMENT: Working environment, either mainnet or test.

  • USE_STREAMABLE_HTTP: Specifies whether your MCP server will run on stdio or streamable-http.

  • PORT: Port where your MCP server will run when using streamable-http.

  • HOST: Host where your MCP server will run when using streamable-http.

Run the project locally

In one terminal window, run the following command: pnpx @modelcontextprotocol/inspector pnpx tsx ./src/index.ts in stdio mode.

Build and run

Run the command: pnpm run build and then: pnpm run start

Deployment

To deploy this MCP server, fork this project into your GitHub account, log in to smithery.ai, and click Publish server. Complete the steps, and once it is deployed, add the required environment variables in settings.


License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.