logo
Free, unlimited AI code reviews that run on commit
git-lrc git-lrc GitHub Install Now We'd appreciate a star git-lrc - Free, unlimited AI code reviews that run on commit | Product Hunt git-lrc - Free, unlimited AI code reviews that run on commit | Product Hunt

near-protocol-interface

A secure interface adhering to the Model Context Protocol (MCP) specification for managing NEAR digital assets and invoking decentralized application logic. This utility allows AI entities to conduct operations such as querying account holdings, cryptographically signing requisite transfers, and provisioning new NEAR network identifiers on behalf of authorized principals.

Author

near-protocol-interface logo

nearai

MIT License

Quick Info

GitHub GitHub Stars 22
NPM Weekly Downloads 0
Tools 1
Last Updated 2026-02-19

Tags

nearaiblockchaincryptocrypto nearainear accountsnearai near

npm version Dev Support Channel

NEAR Context Adapter

This software package represents a server implementation compliant with the Model Context Protocol (MCP) standards, specifically tailored for interaction with the NEAR distributed ledger technology (DLT). It furnishes AI agents and Large Language Models with a secure gateway to execute NEAR ledger operations and manage associated user accounts.

Initial Setup Guide

To rapidly instantiate the near-mcp server using the claude command-line utility:

npm install -g @anthropic-ai/claude-code claude mcp add near-mcp npx @nearai/near-mcp@latest run claude

Alternatively, remote deployment of the MCP host is achievable via Phala Cloud; refer to the documentation found here.

Installation Procedure

near-mcp is fundamentally designed for integration with an MCP-compatible consuming client. Comprehensive details are available within the primary MCP documentation.

Integrating with the claude CLI tool:

bash claude mcp add near-mcp npx @nearai/near-mcp@latest run

Configuration for the Claude desktop application via JSON manifest:

{ "mcpServers": { "near-mcp": { "command": "npx", "args": ["-y", "@nearai/near-mcp@latest", "run"], "env": {} } } }

Configuration steps for the goose orchestration utility:

┌ goose-configure │ ◇ What configuration subject applies? │ Add Extension │ ◇ What extension category? │ Command-line Extension │ ◇ What identifier should the extension possess? │ near-mcp │ ◇ What executable instruction should be invoked? │ npx @nearai/near-mcp@latest run │ ◇ Specify the operational timeout for this utility (in seconds): │ 60 │ ◇ Are environment variables to be supplied? │ No │ └ Extension 'near-mcp' successfully registered

Alternatively, direct global installation or invocation via npx is supported:

bash

Global installation

npm install -g @nearai/near-mcp@latest

Direct invocation via npx

npx @nearai/near-mcp@latest run

Available Functionalities

Consult TOOLS.md for a comprehensive inventory of supported operations and parameter definitions.

AI Model Interoperability

This module is engineered for utilization alongside artificial intelligence systems that conform to the Model Context Protocol. It facilitates AI assistants in the following capacities:

  1. Administration of NEAR accounts on behalf of end-users.
  2. Verification of current account balances and network status.
  3. Digitally signing and broadcasting transactional data.
  4. Provisioning new on-chain identities and managing associated cryptographic credentials.
  5. Inspection and execution of deployed smart contract routines.

Security Posture

  • This specific MCP instance is intended exclusively for local execution environments. Confidential private keys are stored within an unencrypted repository co-located with the MCP host process.
  • The controlling AI models must be architecturally prevented from viewing account credentials, with one specific exception: the import_account function permits the model to integrate an account via a provided private key, necessitating the user to supply this secret information directly to the model.

Community Engagement

We actively solicit contributions to the NEAR MCP host software! Please review the guidelines outlined in CONTRIBUTING.md for detailed submission protocols.

Defect Reporting

Should you identify a software defect or propose an enhancement, kindly initiate a new issue within the official GitHub repository.

See Also

`