Requirements:

• Claude Pro account (or higher tier)

Claude Remote Server Connection

To connect Claude (Web or Desktop) to this MCP server:

📺 Watch the setup tutorial - Click the image below to watch a step-by-step video guide:

📺 Already have your Alpha Vantage MCP server set up? Below are a few examples of Claude performing various stock analysis & charting tasks:

Query Param Option (Recommended):

1. Go to claude.ai/settings/connectors (Web) or Settings → Connectors (Desktop)
2. Click "Add Custom Connector"
3. Add the MCP server URL with your API key: https://mcp.alphavantage.co/mcp?apikey=YOUR_API_KEY (replace YOUR_API_KEY with your actual Alpha Vantage API key)
4. Click "Connect"

OAuth Option:

1. Go to claude.ai/settings/connectors (Web) or Settings → Connectors (Desktop)
2. Click "Add Custom Connector"
3. Add the MCP server URL: https://mcp.alphavantage.co/mcp
4. Click "Connect"
5. Enter your Alpha Vantage API token
6. Click "Authorize Access"

Claude Local Server Connection

See Claude Desktop MCP docs for more info.

Install uv (a modern Python package and project manager):

curl -LsSf https://astral.sh/uv/install.sh | sh

Open Claude Desktop developer settings and edit your claude_desktop_config.json file to add the following configuration:

{
  "mcpServers": {
    "alphavantage": {
      "command": "uvx",
      "args": ["marketdata-mcp-server", "YOUR_API_KEY"]
    }
  }
}

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

Requirements:

• ChatGPT Plus account (or higher tier)
• Developer mode enabled (beta feature) - see OpenAI's Developer Mode documentation

Setup:

1. Go to ChatGPT Settings → Apps
2. Click on Advanced settings → enable Developer mode
3. Return to the Apps submenu, then click Create app.
4. MCP Server URL: https://mcp.alphavantage.co/mcp?apikey=YOUR_API_KEY (replace YOUR_API_KEY with your actual Alpha Vantage API key).
5. Authentication: No authentication

Note: When you return to your conversation, your UI will clearly indicate that you are in Developer mode. ChatGPT will retain memory of any previous remarks made within the same chat session; however, these sessions may not be persisted, meaning the conversation generated while Developer mode is active may not be saved or accessible once closed.

Troubleshooting for ChatGPT Plus Accounts

While Developer mode is available for both Plus and Pro accounts, on Plus accounts, first-time MCP tool execution may exhibit inconsistent behavior. This primarily affects large data payloads returned from the MCP server. If you notice that MCP calls abruptly terminate without returning any JSON response, try the following:

1. Confirm Connection

   • Confirm the connection by prompting:

     List 3 functions available in Alpha Vantage MCP server.
     

   • You should see endpoints like TIME_SERIES_DAILY and TIME_SERIES_INTRADAY.

2. Force JSON Passthrough with simple payload

   • Use the following prompts in sequence to prime ChatGPT to surface the full payload:

     Run ADD_TWO_NUMBERS from Alpha Vantage MCP Server using 3 and 6, and then show the JSON response.
     

   • If successful, ChatGPT should ask you to Confirm or Deny the call. Click Confirm.
   • If it doesn't, use the 🔄 "Try Again..." button and see if you get the Confirm/Deny prompt.
   • If this doesn't work, ask ChatGPT the following, and if it starts Thinking make it Skip, so that it doesn't try the call again and instead outputs its conversational response.

     Why aren't you receiving a JSON response?
     

   • Then, regardless of the response, prompt ChatGPT

     Explicitly request a return value from the tool response to verify what the server outputs.
     

   • Typically, this gets ChatGPT to ask you to Confirm or Deny the call. Click Confirm.
   • Sometimes, ChatGPT will insist it has no active connection to Alpha Vantage MCP Server. Just tell it to "try again".
   • If payloads still fail to appear, start a New Chat and try Step #2 again.

3. Force JSON Passthrough with large payload

   • Once step #2 has been successful, repeat step #2 with the larger call:

     Run TIME_SERIES_DAILY using NVDA, and then show the JSON response.
     

   • Use the fallback steps from Step #2 if ChatGPT does not ask you to Confirm or Deny the call.
   • Once you observe ChatGPT successfully return part of the raw JSON payload from the larger call, you can typically continue on with more conversational queries.

Requirements:

• Mac with Apple Silicon (or dedicated Mac mini)
• LLM API Key (OpenAI used in demo)
• Messaging Platform Bot Token (Telegram used in demo)

📺 Watch the setup tutorial (click image below)

• View the detailed installation guide from GitHub Gist
• For additional support links, please check the video description on YouTube.

Requirements:

• OpenAI account with access to Agent Builder (e.g., billing details added, organization verified)

📺 Watch the setup tutorial. (Click image below.)

OpenAI Agent Builder Connection

To connect OpenAI Agent Builder to this MCP server:

1. Do not add a separate MCP Server object. Instead, click on your agent in OpenAI Agent Builder to add the MCP Server directly to the agent.
2. Find the Tools section of the agent and click the + button to add a new tool
3. Select MCP Server
4. Click "+ Server" in the upper right of the "Add MCP server" menu
5. Configure the MCP server with the following settings:
   • URL: https://mcp.alphavantage.co/mcp
   • Label: Alpha Vantage MCP Server (or any name you prefer)
   • Description: Financial market data and technical indicators (or any description you prefer)
   • Authentication: Select "Access token / API key" (should be default)
   • Access Token: Enter your Alpha Vantage API key
6. Click Connect
7. If successful, the next prompt will display all the tools. Scroll to the bottom and click Add.

Recommended Agent Instructions:

Add the following instruction to your agent's configuration to optimize performance:

You are a helpful financial agent with access to market data through Alpha Vantage MCP Server.

IMPORTANT: Alpha Vantage functions are accessed via wrapper tools:
- Use TOOL_LIST to see available functions (TIME_SERIES_DAILY, RSI, COMPANY_OVERVIEW, etc.)
- Use TOOL_CALL with the format: TOOL_CALL(tool_name="FUNCTION_NAME", arguments={...})
- Example: TOOL_CALL(tool_name="TIME_SERIES_DAILY", arguments={"symbol": "AAPL", "outputsize": "compact"})

To use the Alpha Vantage MCP server with OpenAI Agents SDK, see our example agent that demonstrates:

• Interactive financial analysis agent
• Session management for conversation continuity
• Real-time tool execution with Alpha Vantage data
• Support for both HTTP and stdio MCP connections

The example includes a complete setup guide and configuration templates.

Requirements:

• ChatGPT Plus account (or higher tier)

See OpenAI Codex for more information.

📺 Watch the setup tutorial. (Click image below.) This video additionally provides guidance on handling setup errors and building complete end-to-end applications which are empowered to fetch market data to power dynamic, data-driven visualizations.

• Download the companion .md file from Github Gist.
• For additional support links, please check the video description on YouTube.

Install uv (a modern Python package and project manager):

curl -LsSf https://astral.sh/uv/install.sh | sh

Install Codex CLI v0.34 or later to avoid compatibility issues.

# Install (if not already installed)
npm install -g @openai/codex

# Or update to the latest version
npm update -g @openai/codex

# Verify installation and version
codex --version

Add the following configuration to your Codex MCP server settings by editing ~/.codex/config.toml:

[mcp_servers.alphavantage]
command = "uvx"
args = ["marketdata-mcp-server", "YOUR_API_KEY"]

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

Run codex in your terminal from your project directory. First-time users will be guided through additional prompts.

Then connect with:

/mcp

Requirements:

• Claude Pro account (or higher tier)

See Claude Code MCP docs for more information.

Run one of the following commands, depending on whether you’re connecting remotely or running the server locally.

Claude Code Remote Server Connection

claude mcp add -t http alphavantage https://mcp.alphavantage.co/mcp?apikey=YOUR_API_KEY

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

Claude Code Local Server Connection

First, install uv (a modern Python package and project manager):

curl -LsSf https://astral.sh/uv/install.sh | sh

Then run:

claude mcp add alphavantage -- uvx marketdata-mcp-server YOUR_API_KEY

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

Run claude in your terminal from your project directory.

Then connect with:

/mcp

See VS Code MCP docs for more info.

Create .vscode/mcp.json (your VS Code MCP config file) in your workspace, and paste into it one of the following configurations, depending on whether you’re connecting remotely or running the server locally.

VS Code Remote Server Connection

Paste the following into .vscode/mcp.json:

{
  "servers": {
    "alphavantage": {
      "type": "http",
      "url": "https://mcp.alphavantage.co/mcp?apikey=YOUR_API_KEY"
    }
  }
}

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

VS Code Local Server Connection

First, install uv (a modern Python package and project manager):

curl -LsSf https://astral.sh/uv/install.sh | sh

Then, paste the following into .vscode/mcp.json:

{
  "servers": {
    "alphavantage": {
      "type": "stdio",
      "command": "uvx",
      "args": ["marketdata-mcp-server", "YOUR_API_KEY"]
    }
  }
}

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

Open the Chat view and select Agent mode

See Cursor MCP docs for more information.

Paste one of the following configurations into your Cursor ~/.cursor/mcp.json file, depending on whether you’re connecting remotely or running the server locally. You may also install in a specific project by creating .cursor/mcp.json in your project folder.

Cursor Remote Server Connection

Configure Cursor by editing ~/.cursor/mcp.json:

{
  "mcpServers": {
    "alphavantage": {
      "url": "https://mcp.alphavantage.co/mcp?apikey=YOUR_API_KEY"
    }
  }
}

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

Cursor Local Server Connection

First, install uv (a modern Python package and project manager):

curl -LsSf https://astral.sh/uv/install.sh | sh

Then, paste the following into your Cursor ~/.cursor/mcp.json file:

{
  "mcpServers": {
    "alphavantage": {
      "command": "uvx",
      "args": ["marketdata-mcp-server", "YOUR_API_KEY"]
    }
  }
}

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

See Gemini CLI Configuration for more information.

Run one of the following commands, depending on whether you’re connecting remotely or running the server locally.

Gemini CLI Remote Server Connection (Recommended):

gemini mcp add -t http alphavantage https://mcp.alphavantage.co/mcp?apikey=YOUR_API_KEY

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

Gemini CLI Local Server Connection:

Install uv (a modern Python package and project manager):

curl -LsSf https://astral.sh/uv/install.sh | sh

Then run:

gemini mcp add alphavantage uvx marketdata-mcp-server YOUR_API_KEY

Replace YOUR_API_KEY with your actual Alpha Vantage API key.

Manual Configuration:

1. Open the Gemini CLI settings file. The location is ~/.gemini/settings.json (where ~ is your home directory).
2. Add the following to the mcpServers object in your settings.json file (replace YOUR_API_KEY with your actual Alpha Vantage API key):

{
  "mcpServers": {
    "alphavantage": {
      "httpUrl": "https://mcp.alphavantage.co/mcp?apikey=YOUR_API_KEY"
    }
  }
}

Or, for a local server (replace YOUR_API_KEY with your actual Alpha Vantage API key):

{
  "mcpServers": {
    "alphavantage": {
      "command": "uvx",
      "args": ["marketdata-mcp-server", "YOUR_API_KEY"]
    }
  }
}

If the mcpServers object does not exist, create it.