NGA Wolf Watcher

NGA reply watcher with multi-channel notifications
Watch selected NGA authors or thread authors, then push new replies to Feishu, WeChat, DingTalk, email, or WxPusher.

中文说明 · Latest Release · Installation · Configuration

Quick Start

Windows

1. Open Releases.
2. Download and run nga-wolf-watcher-vX.Y.Z-windows-x86_64-setup.exe.
3. Or download nga-wolf-watcher-vX.Y.Z-windows-x86_64-portable.zip, extract it, and run NGA-Wolf-Watcher.exe.
4. Fill NGA Cookie, then add message channels, NGA users/threads, and listen rules.
5. Before the first long-running watch, mark existing replies as seen to avoid pushing old history in bulk.

The Windows setup installer supports Simplified Chinese and English. Setup, portable, and older onefile builds use the same data directory: %LOCALAPPDATA%\NGA Wolf Watcher\.

macOS

Download nga-wolf-watcher-vX.Y.Z-macos-arm64-experimental.dmg from Releases, open it, then drag NGA-Wolf-Watcher.app to Applications. A .zip package is also published as a fallback.

The macOS build is currently experimental and unsigned. If macOS blocks the first launch, right-click the app and choose Open, or allow it in System Settings. Local config, state, and logs are stored under ~/.nga_wolf_watcher/.

Linux

curl -fsSL https://github.com/huangbwww/nga-wolf-watcher/releases/latest/download/install-linux.sh | sudo bash

Common commands after installation:

sudo ngawolf config
sudo ngawolf check
sudo ngawolf mark-seen
sudo ngawolf test-send
sudo ngawolf start
sudo ngawolf logs -f

If GitHub downloads time out, pin a version and pass a mirror to the installer:

curl -fsSL https://ghfast.top/https://github.com/huangbwww/nga-wolf-watcher/releases/download/vX.Y.Z/install-linux.sh \
  | sudo NGAWOLF_VERSION=vX.Y.Z NGAWOLF_GITHUB_PROXY=https://ghfast.top bash

ghfast.top is not an official project service. Replace it with another working GitHub mirror if needed. More Linux, source, pip mirror, and non-systemd notes are in Installation And Runtime.

How To Copy NGA Cookie

The Cookie is your NGA login credential. The watcher uses it to read pages your account can access.

1. Log in to https://bbs.nga.cn/ in a browser.
2. Open any bbs.nga.cn page.
3. Press F12, then open Network.
4. Keep DevTools open and refresh the page.
5. Select a request to bbs.nga.cn, such as thread.php or read.php.
6. In Headers, find Request Headers.
7. Copy the full Cookie value and paste it into NGA Cookie.

Do not post your Cookie in issues, chats, screenshots, logs, or release files. Empty fetch results or JSON parse errors usually mean the Cookie expired, was copied incompletely, or came from the wrong domain.

Push Channels

Channels, NGA resources, and listen rules are separate: channels store bot accounts, the resource library stores users and thread IDs, and listen rules decide what to watch and where to send. See Configuration, Channels, And Commands.

Common Commands

/start
/setting
/history_r 150058 5
/pack_r 150058 5
/history_t 45974302 10
/pack_t 45974302 10

By default, 150058 is the wolf uid and 45974302 is the wolf thread id. history sends recent replies directly; pack sends packed text or a .txt file. See Commands for channel-specific behavior.

Local AI Agent

AI is disabled by default and is not required for normal NGA watching or pushing. When enabled, the app can pass Feishu/WeChat/DingTalk messages, new NGA replies, and local context to a local Codex, Claude Code, CodeWhale, or custom command.

Typical uses:

• Save watched replies and images into a local AI workspace.
• Ask your local agent from a group with /ai.
• Automatically analyze new replies or run scheduled market-session analysis.
• Keep personal context in context/watchlist.md, context/positions.json, and context/notes.md.

AI output is for information organization, risk notes, and discussion only. It is not investment advice and the app never trades automatically. See Local AI Agent Enhancement.

Stock Dashboard And Bull Strategy

Since v1.5.0, the Windows desktop app includes a stock dashboard and single-stock workbench inspired by bull stock calculator ideas shared in the NGA community. The feature integrates calculator experience shared by Atanvardo_1, Nami, Xhox, and other contributors, then adapts it to this app's local AI Agent context, position list, and focus-watch list.

The dashboard supports watchlists, focus-watch stocks, positions, custom groups, code/name search, fuzzy search, CSV import/export, clearing the list, drag ordering, and 3-second quote refreshes. The market strip includes Shanghai, Shenzhen, ChiNext, and STAR 50 indexes; clicking an index opens the yellow/white intraday line view.

The single-stock workbench supports fast stock switching, intraday/K-line charts, volume, MA/BOLL overlays, swing high/low, Fibonacci retracement, intraday pressure/support, position profit and loss, and bull strategy signals. Swing high/low values are auto-filled from daily K-line and MACD swing logic when possible, but can still be edited manually. Strategy output is for observation and review only. It is not investment advice.

Default Paths

The generated config file includes comments and examples. After manual edits, run ngawolf check or python ngawolf_cli.py check.

Documentation

Run From Source

python -m pip install -r requirements.txt
python .\nga_wolf_webgui.py

Linux CLI:

python -m pip install -r requirements-linux.txt
python ngawolf_cli.py init
python ngawolf_cli.py run

More source runtime, pywebview build, BAT usage, and release packaging notes are in Installation And Runtime.

Feedback

Please open an Issue for bugs, usage problems, or feature ideas.

Before contributing changes, see Contributing. For sensitive reports, see Security Policy.

License

This project is licensed under the MIT License.

Disclaimer

This project is for personal technical research and learning. You are responsible for your own use.

You must confirm and follow the rules of NGA, Feishu, WeChat, DingTalk, email providers, your organization, and applicable laws. Automated access, message forwarding, Cookie usage, frequent polling, and bot interaction may cause account restrictions, bans, data exposure, service interruption, or other unpredictable consequences. The author provides no warranty and is not liable for losses, disputes, account penalties, legal consequences, or third-party claims caused by using, modifying, distributing, or deploying this project.