ThetaDataDxEvery ThetaData endpoint, however you work.

Historical, streaming, and bulk market data through whatever fits how you build: a REST API, four native SDKs, or an MCP server for your AI.

Quickstart

Browse the REST API

View on GitHub

REST API

Every endpoint over plain HTTP. Curl it from the shell or call it from any stack. Typed JSON, CSV, or NDJSON responses.

Streaming

Live quotes, trades, and open interest over WebSocket, delivered as typed events through a callback you register once.

MCP for AI

Point Claude or any LLM at ThetaData and ask for an options chain in plain English. A tool per endpoint, no code to write.

Native SDKs

Rust, Python, TypeScript, and C++ over one identical surface. Zero-copy decode, the same method names in every language.

Self-host Server

Run the HTTP + WebSocket gateway yourself on the v3 route surface. Existing scripts point at it unchanged.

A quote in seconds

The latest NBBO quote for AAPL, four ways. Pick the one that fits your stack. Same data, same fields, from a free or paid subscription.

cURLPythonTypeScriptRust

bash

# Start the server once: thetadatadx-server --api-key "$THETADATA_API_KEY" &
curl 'http://127.0.0.1:25503/v3/stock/snapshot/quote?symbol=AAPL'

python

from thetadatadx import Client

client = Client(api_key="your_api_key")
[q] = client.market_data.stock_snapshot_quote(["AAPL"])
print(q.bid, q.ask)

typescript

import { Client } from 'thetadatadx-ts';

const client = await Client.connectWith({ apiKey: 'your_api_key' });
const [q] = await client.marketData.stockSnapshotQuote(['AAPL']);
console.log(q.bid, q.ask);

rust

let client = thetadatadx::Client::builder().api_key("your_api_key").connect().await?;
let rows = client.market_data().stock_snapshot_quote(&["AAPL"]).await?;
println!("{} {}", rows[0].bid, rows[0].ask);

What are you building?

Goal	Start here
A research notebook	Python SDK: install, authenticate, pull history into pandas or polars.
A live signal or dashboard	Streaming: typed quote, trade, and open-interest events over WebSocket.
An AI assistant or agent	MCP for AI: give any LLM client a tool per endpoint, no code.
A backend or existing tool	REST API or the self-host server: plain HTTP on the v3 routes.

Install

The SDK ships as per-language packages at 0.1.0: thetadatadx-rs (crates.io), thetadatadx-py (PyPI), and thetadatadx-ts (npm). Install directly below.