KingshotAuto

A Windows desktop automation tool for the mobile strategy game Kingshot. It runs the game inside the LDPlayer Android emulator and automates in-game tasks using computer vision (template matching) and OCR.

Status: early / not fully working — contributions welcome. KingshotAuto was previously a paid, closed-source product. It is now MIT-licensed open source. Not every automation task works reliably yet. It is published so the community can fix, improve, and extend it. See Known issues / roadmap and CONTRIBUTING.md — beginners are very welcome.

---

Features

KingshotAuto drives the game through a set of pluggable automation tasks. Each task uses screen captures from the emulator plus template matching / OCR to find and click the right UI elements. Current task modules include:

• Farming — gather resources on the world map
• Auto Hunt — attack hunting / beast targets
• Auto Rally — start and join rallies
• Auto Shield — keep your city shield active
• Auto Heal — heal wounded troops
• Auto Build — queue building upgrades
• Troop Training — train troops
• Auto Technology & Alliance Technology — research / alliance tech
• Auto Claim Hero — claim hero rewards
• Alliance Help — send alliance help
• Conquest Collect — collect conquest rewards
• Collect VIP — claim VIP rewards
• Claim Mail & Claim Missions — collect mail and mission rewards
• Resident Welcome — welcome residents
• Account Detection / Change Account — multi-account switching
• Startup / Recovery — bring the game to a known state and recover from unexpected UI

Additional capabilities:

• Multi-instance support — run multiple emulator instances / accounts with per-instance throttling
• Cycle management — repeat task batches on a randomized schedule
• Computer-vision pipeline — multi-scale template matching with grayscale and HSV color spaces
• OCR via Tesseract for reading numbers and text on screen

Screenshots

Screenshots coming soon — contributions welcome! (PRs adding screenshots of the UI and a running cycle are appreciated.)

Prerequisites

You will need the following on a 64-bit Windows machine:

• Windows 10 or 11 (x64)
• .NET 10 SDK — https://dotnet.microsoft.com/download/dotnet/10.0
• Git with Git LFS — the native OCR DLLs under runtimes/win-x64/native/ and tools/VC_redist.x64.exe are stored via Git LFS. Without git-lfs they clone as tiny pointer files and Tesseract OCR fails at runtime. Install Git LFS once with git lfs install before cloning (or run git lfs pull afterwards).
• LDPlayer 9 (64-bit) emulator — https://www.ldplayer.net/
• Microsoft Visual C++ 2015–2022 x64 Redistributable — required by the OpenCV / Tesseract native libraries
• Administrator privileges — the app must run as Administrator to control the emulator over ADB

Quick Start

# 1. Make sure Git LFS is installed (only needed once per machine)
git lfs install

# 2. Clone the repository (LFS-backed native files download automatically)
git clone https://github.com/KingshotAuto/Kingshot-bot.git
cd KingshotAuto

# If you cloned before installing Git LFS, pull the real binaries now:
git lfs pull

# 3. Restore dependencies
dotnet restore BotApp.sln

# 4. Build (Debug)
dotnet build Bot.GUI/Bot.GUI.csproj -c Debug

# 5. Run (launch your terminal as Administrator)
dotnet run --project Bot.GUI/Bot.GUI.csproj

Run as Administrator. Controlling LDPlayer over ADB requires elevated privileges. Start your terminal (or the built app) as Administrator, otherwise the emulator/ADB connection will fail.

Maintainers can produce a portable packaged build:

# Produces a portable zip in dist/ ; add -SkipInstaller to skip the NSIS installer
./Build.ps1 -Version 0.1.0

Configuration

• On first run, the app auto-creates its configuration at configs/config.json. If that file is missing it falls back to a built-in default, so you can start without any manual setup.
• A documented example of every option lives at configs/default_config.example.json — copy it to configs/config.json if you want to edit settings by hand.
• The LDPlayer install path defaults to C:\LDPlayer\LDPlayer9. The app tries to auto-detect it, and you can also set it from the in-app Browse button.
• Most settings (which accounts to run, which tasks are enabled, per-task options, and cycle timing) are configurable from the UI.

How it works

1. Emulator control — KingshotAuto talks to LDPlayer over ADB (via AdvancedSharpAdbClient) to launch instances, capture screenshots, and send taps/swipes.
2. Computer vision — each task captures the emulator screen and uses OpenCVSharp4 template matching to locate buttons, icons, and targets, then clicks them. Matching uses search areas and tuned thresholds, with grayscale and HSV color spaces and multiple scales for robustness.
3. OCR — Tesseract reads on-screen numbers/text where template matching is not enough.
4. Orchestration — a cycle manager runs the enabled tasks per account on a randomized schedule, with retry and recovery logic for flaky UI/connectivity.

Template images for the CV pipeline live in templates/images/, organized into one folder per task. Each task maps to its folder via GetImageFolderName().

Project structure

KingshotAuto/
├── BotApp.sln                 # Solution file
├── Bot.Core/                  # Core automation logic
│   ├── Tasks/
│   │   ├── TaskManager.cs      # Registers and runs tasks
│   │   └── Modules/            # Individual automation tasks
│   ├── Services/               # Cycle management, instance control, etc.
│   ├── ImageDetection/         # Template matching (OpenCV)
│   ├── LDPlayer/               # Emulator / ADB control
│   ├── Config/                 # Configuration management
│   ├── Models/                 # Data models (incl. TaskType enum)
│   └── Utils/                  # OCR, locators, helpers
├── Bot.GUI/                    # WPF UI (MVVM)
│   ├── ViewModels/
│   ├── Views/
│   └── Converters/
├── Bot.Core.Tests/            # Tests
├── configs/                    # Config + documented example
├── templates/images/          # CV template images (one folder per task)
└── docs/                       # ARCHITECTURE.md, etc.

Contributing

Contributions of all sizes are welcome — bug fixes, new/working tasks, tests, docs, and screenshots. Please read CONTRIBUTING.md for dev setup, how to add a new automation task, how to capture template images, and the PR workflow.

Use Issues to report bugs and Discussions to ask questions and coordinate work.

Known issues / roadmap

This project is honest about its rough edges — these are great places to start contributing:

• Not all tasks work reliably. Several task modules need fixing or tuning.
• Resolution sensitivity. Template images were captured at a specific emulator resolution, so matching can fail on other resolutions/DPI settings. Making matching resolution-independent is a high-value contribution.
• Flaky ADB / emulator connectivity. Connection handling can be brittle; improvements to reconnection and recovery are welcome.
• Test coverage is thin. More unit/integration tests would help a lot.

If you want to help, see CONTRIBUTING.md and open/claim an issue.

Privacy

KingshotAuto runs entirely on your machine. No telemetry, no analytics, no license server, no phone-home. The original licensing/activation, analytics, and auto-update systems have been completely removed — a build contacts no servers and sends no data anywhere.

Disclaimer

This software automates a third-party game and may violate that game's Terms of Service. Using automation tools can put your account at risk of suspension or banning. Use at your own risk. KingshotAuto is provided "as is", without warranty of any kind (see the MIT License). The authors and contributors are not affiliated with the makers of Kingshot and accept no liability for how you use this software.

License

Released under the MIT License. See LICENSE.

Copyright (c) KingshotAuto