← BASELOAD English

Stock Trading APIs Compared — KIS, Kiwoom, Toss

June 27, 2026 · 36 min read 한국어 원문

A broker’s open API is “the official channel to control your brokerage account with code.” Once you hook up a stock trading API, you can receive market data in real time and place buy/sell orders automatically, or pull data without ever opening a chart window to run quant analysis or market-data bots. Even though they all carry the same “open API” label, every broker connects in a completely different way, with different authentication procedures and different operating systems (OS) they run on. Pick the wrong one and you end up “trying to deploy it to a Linux server, only to give up because it’s Windows-only.” This article compares Korea Investment & Securities, Kiwoom, Toss Securities, LS, and Daishin from an automated-trading and market-data standpoint, as a practical guide to choosing the API that fits your environment.

First, let me be clear: this article compares APIs as developer tools, not as investment advice. It does not recommend buying or selling any particular security, and all responsibility for investment decisions and the operation of automated trading rests with you. This is not investment advice.

Photo by Tran Mau Tri Tam ✪ on Unsplash

The Key Divide: 1st-Gen OCX vs 2nd-Gen REST

The very first thing that separates broker APIs is the “technology generation.” This single line determines all the other differences.

1st generation — OCX/COM approach. Kiwoom OpenAPI+, Daishin Creon, and LS XingAPI fall into this category. You install a Windows-only component (OCX/COM) on your PC, and the broker’s own program (e.g., Kiwoom’s HeroMoon) has to be running resident in the background for it to work. In other words, it’s a setup where “you keep a Windows PC powered on, and on top of it your trading program talks to the broker’s program.” It’s usually tied to a 32-bit environment and can’t be deployed to a Linux server or the cloud. On the other hand, because it’s been around so long, there’s an abundance of documentation and examples.

2nd generation — REST/WebSocket/gRPC approach. Korea Investment & Securities’ KIS, Toss Securities Open API, LS OpenAPI, and Kiwoom’s new Open API NEXT fall here. You obtain a token over HTTP and call orders/queries with it, so it’s OS-agnostic. You can connect with any language — Python, JavaScript, and so on — and deploy it to a Linux cloud server for 24/7 unattended operation. Real-time market data comes through WebSocket (or a gRPC stream).

To sum up, the core of the choice is this: “If you’re going to run it unattended on the cloud/Linux, go 2nd-gen REST; if you already have Windows-based assets, go 1st-gen.” The whole market is also rapidly shifting toward REST/gRPC, so if you’re starting fresh today, 2nd gen is the default.

Broker-by-Broker Comparison Table

The table below pulls out only the key items relevant to automated trading and market-data retrieval. Fees and exact call limits (requests per minute) change policy frequently, so I won’t state them as fixed. Be sure to check the latest values on each broker’s developer center.

Paper trading and detailed support scope can vary over time, so I’ve marked them “needs checking.” Before you build on it, it’s safest to check each developer center’s documentation.

Who Each Broker Suits

Korea Investment & Securities KIS Developers. Since 2022 it has offered a REST API and real-time WebSocket, and it supports paper trading too. Its official GitHub repository (koreainvestment/open-trading-api) has well-organized Python examples and it’s free, so among Korean individual developers it’s effectively close to a standard. Its biggest strength is that there’s the most documentation, so when you get stuck you can usually solve it by searching. I’d recommend it as the top pick for anyone building their first automated trading system or trying to run market-data bots and quant strategies in Python.

Kiwoom Securities. For a long time it grew the individual automated-trading ecosystem on the 1st-gen OpenAPI+ (OCX/COM). You have to keep HeroMoon running resident on a Windows PC and you’re tied to a 32-bit environment, but in return there’s a vast amount of blogs, books, and examples. In 2026 it’s introducing REST and gRPC with the new ‘Open API NEXT,’ moving into the 2nd generation. It suits people who already have existing Kiwoom-based code or are comfortable with a resource-rich 1st-gen environment, and those who are willing to migrate to the new REST.

Toss Securities. Newly launched in 2026, its Open API uses REST and OAuth2.0 client credentials (server-to-server auth). It supports order automation for domestic (KRX) and US stocks, and you can get started with no separate fees once you complete in-app pre-application and account opening. Its barrier to entry is the lowest, making it a good fit for the mobile generation or anyone who wants to casually try wiring up automated orders.

LS Securities (formerly Ebest) and Daishin Securities. LS offers both the 1st-gen XingAPI (OCX) and the new LS OpenAPI (REST), giving you wide transitional options. Daishin’s Creon is a classic 1st-gen, Windows COM-based offering with a solid existing Creon user base. For anyone who already has assets with that broker, both are natural choices.

Photo by Dimitri Karastelev on Unsplash

What Does the Auth Flow Look Like

2nd-gen REST APIs have almost identical authentication flows. It’s the OAuth2.0 token pattern: “first get a token using your issued keys, then call orders/queries with that token in the header.” Below is KIS-style pseudocode (actual endpoints and field names follow the developer center documentation).

import requests

BASE = "https://openapi.example-broker.com"  # live/paper servers are separate

# 1) Issue an access token with the app key/secret
res = requests.post(f"{BASE}/oauth2/tokenP", json={
    "grant_type": "client_credentials",
    "appkey": APP_KEY,
    "appsecret": APP_SECRET,
})
token = res.json()["access_token"]

# 2) Query the current price with the token in the header
headers = {"authorization": f"Bearer {token}", "appkey": APP_KEY, "appsecret": APP_SECRET}
price = requests.get(f"{BASE}/quotations/price",
                     headers=headers, params={"symbol": "005930"})

# 3) Subscribe to real-time quotes via a separate WebSocket (or gRPC stream)

Toss Securities’ client credentials follow the same skeleton. Tokens have an expiration time, so add caching and auto-reissue logic, and don’t hardcode keys and secrets into your code — keep them in environment variables. 1st-gen OCX/COM doesn’t issue tokens like this; instead it’s a setup where “the program attaches to an HTS login session,” so the very concept of authentication is different.

Selection Guide — So What?

Let me boil it down to three situations.

• Automated-trading beginners / Python market bots and quant: Korea Investment & Securities KIS. It’s free with an overwhelming amount of resources, so even when you get stuck you can solve it by searching. Validate first with paper trading.
• 24/7 unattended operation on the cloud: 2nd-gen REST/gRPC (KIS, Toss, LS OpenAPI, Kiwoom NEXT). Deploy it to a Linux server and you won’t have to keep a PC on. 1st gen requires a resident Windows machine and is unsuitable for this use case.
• Easiest to start / includes US stocks: Toss Securities. With just in-app pre-application, you can quickly wire up OAuth-based order automation.
• If you already have Windows assets: Kiwoom OpenAPI+, Daishin Creon, LS XingAPI. If you already have code that works, there’s no need to force a migration — but do keep an eye on the roadmap toward the new REST.

My view is clear. If you’re starting fresh today, 2nd-gen REST is the default, and among those, KIS offers the best bang for the buck relative to the learning curve. I see OCX/COM as a reasonable choice “only when you already have assets there.”

One final word of warning. Automated trading can produce real losses from a single line of buggy code, overfitting (strategies that only look good in backtests), or sudden market swings — and that responsibility rests entirely with you. Be sure to validate thoroughly in paper trading, build in safeguards like order size and daily loss limits in your code, and start with a small amount. To emphasize again: this article is merely a comparison of developer tools and is not investment advice.