TradingAgents UI

中文说明

A local web app for TradingAgents. Open it like a desktop app, enter your settings, run an analysis, and read the report in the same window.

One-Click Launch

macOS

Create the app once:

./scripts/install-macos-app.sh

Then double-click:

TradingAgents UI.app

It starts the local Streamlit server and opens http://localhost:8501 in a standalone Chrome/Edge app window.

TradingAgents UI.app is macOS-only. On Windows, use the .bat launcher below.

Windows

After the first install, double-click:

scripts\launch-local-webapp.bat

It starts the local server in the background and opens:

http://localhost:8501

For a more app-like workflow, create a desktop shortcut to the .bat file.

All Platforms

trade-ui

From this checkout during local development:

./run.sh

First Install

git clone https://github.com/Kevoyuan/tradingagents-ui.git
cd tradingagents-ui
python3 -m pip install -e .

Then choose your launcher:

• macOS: run ./scripts/install-macos-app.sh, then double-click TradingAgents UI.app
• Windows: double-click scripts\launch-local-webapp.bat
• Linux/other: run trade-ui

How To Use

1. Open the local UI
2. Fill in ticker, date, language, analyst team, model, and API keys in the sidebar
3. Click Run Analysis
4. Open Browse Reports when the analysis is done
5. Read the embedded HTML report, or switch back to Markdown

API keys are saved locally for the next launch. Cloud deployments keep keys session-only.

Built-In TradingAgents Update Check

The app silently checks GitHub once when it opens.

If an update is available, an update icon appears next to the TradingAgents logo in the upper-left sidebar. If the icon is not there, the app did not detect an available update.

Click the icon to install or update TradingAgents from GitHub. Restart the app afterward so already-loaded Python modules refresh cleanly.

Reports

Browse Reports shows embedded HTML by default, so you do not need a separate browser window.

Embedded HTML rendering is local-only. It uses the saved Markdown report and does not call an LLM or spend API tokens.

You can:

• Select a report, switch HTML/Markdown, and copy Markdown from one toolbar
• Jump through sections with the report table of contents
• Use fixed Top/Bottom buttons for long reports
• Read reports in the dark Quant Terminal theme

Historical reports are stored under:

~/.tradingagents/logs/.../reports/

Phone Access

When your computer and phone are on the same Wi-Fi:

trade-ui --lan

or:

./run-lan.sh

Open the printed URL on your phone, for example:

http://192.168.1.23:8501

Cloud Deploy

Deploy to Streamlit Community Cloud if you want access without keeping your computer online:

1. Push this repo to GitHub
2. Create a new Streamlit Community Cloud app
3. Set Main file path to app.py
4. Use Python 3.10 or newer
5. Do not store API keys in code or app secrets
6. Deploy and open the generated URL

Cloud notes:

• Each user enters their own API key in the sidebar
• Cloud reports live inside the cloud container and are best for temporary viewing
• Local-only services such as Ollama or localhost LiteLLM are not reachable from Streamlit Cloud

Features

• One-click local launch
• macOS standalone app window
• Windows double-click launcher
• Sidebar setup for ticker, date, language, analyst team, model, and API keys
• Built-in TradingAgents GitHub update check
• Live agent progress, messages, tool calls, token counts, and timing
• Embedded historical HTML reports
• One-click Markdown copy
• HTML/Markdown view switch
• Report table of contents plus Top/Bottom navigation
• Local preference and API key persistence
• Same-Wi-Fi phone access

More Screenshots

Live Analysis Monitor

Report Viewer

Report History

Developer Notes

UI entrypoint priority:

1. TRADINGAGENTS_UI_APP_PATH
2. ./app.py in the current directory
3. Packaged fallback app.py

For a local TradingAgents checkout:

export TRADINGAGENTS_DIR=/path/to/tradingagents

Project layout:

tradingagents-ui/
├── app.py
├── ui_config.py
├── ui_styles.py
├── ui_panels.py
├── preferences.py
├── scripts/
│   ├── install-macos-app.sh
│   ├── launch-local-webapp.sh
│   └── launch-local-webapp.bat
├── tools/
│   └── baoyu-markdown-to-html/
├── trade_ui/
│   └── cli.py
├── pyproject.toml
├── run.sh
└── run-lan.sh