Skip to content

Real-time data for OpenAI agents

Add live crypto and stock market data to agents built with the OpenAI Agents SDK via MCP or plain x402 HTTP calls.

  • Mount the MCP server with `MCPServerStreamableHttp`.
  • Or call any endpoint over HTTP and pay with x402 from the agent's own wallet.
  • Pay-per-time keys for agents that run all day.
Python
from agents import Agent
from agents.mcp import MCPServerStreamableHttp
async with MCPServerStreamableHttp(params={"url": "https://mcp.realtimeapi.ai/mcp"}) as rt:
    agent = Agent(name="Analyst", mcp_servers=[rt])
Or in Claude Code
claude mcp add --transport http realtimeapi https://mcp.realtimeapi.ai/mcp

Feeds for this use case

Frequently asked questions

Do agents built with this framework need an API key?

No. Agents can pay per request in USDC with x402, or start with a free key (1,000 calls a month across every feed).

What does a call cost?

Each endpoint has a provider-set price, $0.008 to $0.10 per call on these feeds. Only successful responses are billed.

Get a free key

1,000 calls a month across every feed. Who is setting it up?

No key needed at all: agents can pay per call in USDC with x402.