OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
MCP & Tooling

Arcade MCP

Arcade MCP is an open-source Python framework and CLI for building Model Context Protocol servers and the tools they expose to AI agents. Developers can define tools with decorators, run servers over stdio or HTTP+SSE, and use Arcade Cloud hosting or a standalone MCP client setup.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Arcade MCP — OpenViST project profile in MCP & Tooling.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Arcade MCP when an agent needs custom MCP tools—for example, to access an internal API or a domain-specific integration—and you want to define and run those tools in Python.

Who it is for

  • Python developers creating custom capabilities for AI agents.
  • Teams integrating internal APIs, domain-specific services or custom OAuth providers through MCP.
  • Developers who need to run their own MCP server locally or deploy and manage one through Arcade Cloud.

Documented uses

  • Create an MCP server and scaffold a Python project with example tools.
  • Expose custom operations as agent-callable MCP tools.
  • Connect tools to services that require OAuth scopes, API keys or other secrets.
  • Run a server through an MCP client over stdio or HTTP+SSE.
  • Deploy a server to Arcade Cloud and use its documented logs, status and dashboard commands.
Before you choose

Requirements & limits

The project is open source under the MIT license. The documented source-install path requires Python 3.10 or later and uv.

  • Choose stdio or HTTP+SSE based on the MCP client and integration path; the README gives client configuration examples for Claude Desktop, Cursor and VS Code.
  • Arcade Cloud and standalone hosting have different responsibilities: standalone users provide access tokens for authenticated tools and protect production HTTP endpoints themselves.
  • The README describes Arcade Cloud handling OAuth authorization, token refresh and per-call scoping for declared tool requirements; generic OAuth providers require registering an OAuth app in the Arcade Dashboard.
  • The project includes an arcade evals command for evaluating tool-call accuracy against real LLMs.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Scaffold and run a local MCP server

Create a starter server and run it over stdio or HTTP+SSE for a compatible MCP client.

Before you start

  • The arcade-mcp CLI, installable with uv.
  • A Python project scaffold generated by the CLI.

Steps

  1. Install the CLI with the command below.

    Documented command
    uv tool install arcade-mcp

  2. Run arcade new my_server and enter my_server/src/my_server.

  3. Run the command below for stdio, or the command below for HTTP+SSE.

    Documented command
    uv run server.py
    Documented command
    uv run server.py http

  4. Use arcade configure with the documented client name and transport to configure Claude Desktop, Cursor or VS Code.

Expected outcome

A scaffolded MCP server can be run locally using one of the documented transports and configured for a listed client.

Guide limitations

  • The guide documents scaffolding, running and client configuration, but not implementation of a particular application integration.
  • Authenticated tools in standalone mode require the developer to supply access tokens; production HTTP endpoints should be protected with Resource Server Auth.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
12 October 2026
Original website title
ArcadeAI/arcade-mcp · Official homepage title · retained in its source language
How we review sourcesSuggest a correction