Setup Guide for OpenAI Playground

OpenAI Playground Configuration Guide

Log into the OpenAI Playground and pick a compatible model. Currently, MCP is supported on newer models (e.g. gpt-4.1 or later) via the Playground interface. Once signed in, create a playground session with the desired model

Step 1: Open the Tools (Plugins) panel. In the Playground UI, there is a Tools section or sidebar. Click to expand it. Here you can add integrations or tools that the model can call. Choose MCP Server

Step 2: Add an MCP server integration. Click the “Create” button (or similar) in the Tools panel and choose the provider for your MCP server.Authenticate and configure the server. After selecting a provider, you will be prompted for credentials or configuration. Choose Headers and add “Authorization” : “Your Bearer Token”. Once the credentials are valid, the MCP server is added to your Playground session and will appear in the Tools list.

Step 3: Enable the tool. In the Tools panel you can toggle the new MCP server on or off. Make sure it is enabled for this chat. You can often expand it to see additional settings (for example, a checkbox for require_approval or a list of allowed tools).

1. Usage

Once connected you can chat, and vibe code your agent

web3_insights_analysis	Technical analysis for cryptocurrencies	crypto_symbol (e.g., "BTC")	Perform short term analysis for BTC
web2_insights_analysis	High-conviction signals for stocks, commodities and FX	asset_symbol (e.g., "AAPL")	Perform short term analysis for Gold

You can simply use prompt query mode to trade if you are performing low frequency trades