Building AI Trading Agents with Browser Use for Interactive Brokers

Learn to develop AI trading agents using the open-source Browser Use project in this 38-minute tutorial video. Explore practical demonstrations of AI agents performing various trading tasks, including running stock scanners, placing limit orders, building watchlists from Stocktwits feeds, and conducting financial research using Perplexity Finance. Follow along with detailed code walkthroughs covering Browser Use implementation, environment setup, API integrations, and connecting to Interactive Brokers via Docker containers. Gain hands-on experience with both paper and live trading setups, while understanding the technical aspects of making web applications AI-agent compatible. Master advanced features like capturing session activities, implementing custom API calls, and automating trading strategies. Important caution is given regarding the risks of automated trading, especially noting that open-source solutions lack the safety restrictions found in platforms like OpenAI Operator.

Syllabus

0:00 Introduction: Browser Use Project
2:25 Demo: AI Agents in action: running a stock scanner
4:33 Demo: AI Agent places a limit order WARNING: very risky!
5:38 Demo: AI Agent builds a watchlist from a Stocktwits user feed and copy trades
7:45 Demo: AI Agent uses Perplexity Finance as a research tool, extracts Tesla earnings reaction
11:38 Capture an animated gif of what the agent did in its session for monitoring
12:45 Checking out the source code from my Github
14:22 Browser use overview, packages and dependencies
17:22 Code walkthrough of “Hello World” of browser use agent
20:02 Environment variables, OpenAI API Key settings
21:55 Running the agent, capturing a reddit comment
23:10 You can use open source models and give it custom API calls / functions to use
24:30 Connecting it to your brokerage / exchange
24:58 Getting the Interactive Brokers Web API docker container / source code
26:38 Flask App Python code walkthrough
27:58 Bringing up the docker container
28:41 Authenticating, paper trading vs. live trading
30:11 Making web apps useable by AI agents
30:34 Demo: showing the agent how to scan for high dividend yield stocks
34:15 Demo: copying trading a stocktwits user
36:33 Unlike OpenAI Operator, open source doesn't prevent risky transactions