Skip to content

Latest commit

 

History

History
46 lines (27 loc) · 3.91 KB

README_en.md

File metadata and controls

46 lines (27 loc) · 3.91 KB
Raw

| English | 简体中文 | Español |

StarRailCopilot

Star Rail Instant Tea, a bot for Honkai: Star Rail, based on the next generation of ALAS framework.

gui

setting

Features

  • Dungeons: Character Planner, Dungeons, dungeons at double event, echo of war.
  • Rewards: Complete daily quests, do assignment, claim nameless horner rewards.
  • Simulated Universe: Farm SU, farm planers using trailbalze power.
  • AFK Auto Farming: Automatically launch emulators and games, do dungeons and quests at background, keep track of resources through dashboard.
  • Cloud Gaming: [CN only] Run SRC on cloud HSR

Install

Refers to Installation tutorial, including easy-installer manual, use manual, manual installation tutorial.

Why use emulators? If you run a bot on the desktop client, game windows must stay at front. I guess you don't wanna baby-sit there without being able to move the mouse and keyboard while running the bot, so use the emulators.

How's the performance? Lme's 8700k+1080ti using MuMu12 emulator with graphic settings very high gets 40 fps. It shouldn't be a problem to run with maximum graphic settings and 60 fps if you have newer PC specs.

Develpment

Discord https://discord.gg/aJkt3mKDEr

  • Minimap Tracking

  • Development Docs (menu is on sidebar): Alas wiki (in Chinese). However, there is ton of code newly written, it is recommended to read the source code and historical commits.

  • Development Road Map: See pinned issues. Pull Requests are welcomed, just pick the part you interested to work on.

How to add multi-language or multi-server support? Need assets updates, see "Adding a Button" in development docs.

About ALAS

SRC is based on an Azur Lane bot AzurLaneAutoScript. After 3 years of development, ALAS has reached a high degree of completion, but it has also accumulated a lot of shit code that is difficult to change. We hope that Fix these issues on a new project.

  • Update OCR. ALAS has trained multiple models on cnocr==1.2.2, but the dependent mxnet is no longer active, machine learning is developing rapidly, and the speed of new models and the correct rate crushes the old model.
  • Converting setting files into pydantic models. Since the concept of task and scheduler was added, the number of user settings greatly increased. ALAS has built a code generator to implement setting read and update. pydantic will make things more elegantly.
  • Better Assets management. button_extract helps ALAS to easily maintain 4000+ template images, but it has a serious performance issue, and the reminder of multi-server supported is also submerged in a large amount of meaningless logs.
  • Reduced coupling to Azur Lane. The ALAS framework and ALAS GUI have the ability to interface with other games and their not, but the completed MAA plug-in for Arknights and [fgo-py](https: //github.com/hgjazhgj/FGO-py) plug-in under development have found serious coupling problems between ALAS and the Azur Lane game itself.