Skip to content

clash-jev

A Clash Royale bot that turns a phone screenshot into JSON once a second and asks Jev for a strategy, a card, and a square.

Category
Demos & Experiments
Format
—
Published by
Community
Author
bytelabs-oss
Added
2026-09-28
Last verified
2026-09-28

bytelabs-oss/clash-jev

Organization repository on GitHub

View on GitHub

A Clash Royale bot with no trained policy: Jev (TypeSafe System One) makes every decision from the live game state

GitHub stars
33
Forks
12
Primary language
Python
License
mit
Last pushed
Updated Sep 2026

Repo stats from the GitHub API, cached Sep 2026.Homepage

Highlights

  • No policy is trained. OpenCV and a small hand-labeled troop network build the state. Jev only chooses.
  • A play is three requests: strategy, card, then square. Choosing not to play costs one.
  • The state holds measurements only. Derived facts such as a counter list were removed after they moved a card to 98%.
  • Across about 20 matches, a match took 150 to 200 requests and about $0.004.
  • A failed request plays a baseline move that the log marks as not from Jev.

Quickstart

python
state = extract_state(screenshot)
options = get_valid_moves(state)
move = choose_with_jev(state, options)

Watch out

MIT. Automating play is against Supercell's terms. A research project, not affiliated with Supercell or TypeSafe. Needs a real Android device over adb.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

X: Jev driving real Chrome from Grok Bot

Took Gregor’s Ultrafast idea and wired it into Grok Bot. @bot @OpenRouter @typesafeai @gregpr07 Your bots can now use Jev to drive the real Chrome on the machine instead of slow look-and-click. Drop in the API key you already have (OpenRouter or TypeSafe), and it gets going. Show more

Grok
Grok
SpaceXAI
@grok

We’re inviting one Grok Bot user to a Starship launch. You could win the invite by sharing the ways you've integrated @Bot into your work. The invite includes a plus one. How to submit and some examples we love:

Reply

X: Voice-controlled computer use on a Mac

X: A computer assistant that listens

X: Headless Chromium agent

Related terms

Glossary definitions related to this entry.

More like this

A Mineflayer agent for Minecraft Java 1.16.5 where a planner sets the objective and Jev picks the next bounded action from game observations.
Demos & Experiments#community#javascript#gaming
1kGitHub stars
A macOS computer-use loop that OCRs the screen, asks Jev for the next action with typed Choices, and clicks, at about $0.0002 per decision without sending screenshots to a frontier model.
Tools & CLIs#community#python#agents
101GitHub stars
A hands-free macOS voice assistant: whisper.cpp transcribes locally in about 100 ms, then one roughly 250 ms Jev request selects a typed action plus arguments, and code executes it against the Mac.
Tools & CLIs#community#python#voice

From the guides

Original write-ups that draw on this entry.

A game tick is a state plus a legal action list. How Clash Royale, Minecraft, and the earlier Mario, Pokémon, StarCraft, and drone loops ask Jev one step at a time.
Demos & Experiments#gaming#agents#control
Read article
Back to all resources

From the community

Posts from builders shipping with Jev right now.

Follow @typesafeai

Screening agent actions with Jev

Tested TypeSafe’s Jev (no-text, probability-only model) as an AI agent safety monitor. Checking each action first worked well caught most attacks with almost no false blocks, and much faster than Gemini.

Image
Image
Image
Diogo Almeida
Diogo Almeida
TypeSafe AI
@CompleteSkeptic

After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x

Reply

Cua's small System One models

A 706K-parameter form filler

cua open sourced a 706k param model that fills a whole form in one 50ms pass the llm agent doing the same form took 23 turns and 39.6 seconds the specialists are going to eat the generalists from the bottom

Cua
Cua
@trycua

1/ Introducing CUA-S1: a family of System One Models, small, specialized, and built for computer use. Today we're open-sourcing CUA-S1-FORMS, the first in the family: github.com/trycua/cua

Image
Reply

Navigating Neo4j with Jev

Jev 这个 waitlist 还是很给力的,昨天申请,今天就能用上。 给已经拿到 API、但还不知道怎么玩的人整理了一份 Awesome Jev,目前我能确认到的 Jev 项目基本都在这里: 1. jev-ultrafast Browser Use 做的高速浏览器 Agent。Jev Show more

Image
思维怪怪
思维怪怪
@0xLogicrw

前 OpenAI 研究员 Diogo Almeida 创办的 TypeSafe AI 推出新模型 Jev。它有点像一个能读懂自然语言的超级分类器,不生成文本,只返回选项、分数和概率,专门给软件做判断。 普通大模型需要一个 token 一个 token 往外生成,Jev 则可以并行给出多个结果。TypeSafe 还用新的 RLCD

Reply

Reranking 33,047 catalog entries

拿 Jev 做搜索重排,我先泼一盆冷水:单独用,它没打赢向量检索 TypeSafe 的 Jev 这阵子很火,一堆项目拿它做重排。我们在 Agent Skills Hub 的 33,047 条目录上认真测了一次,164 条中英文真实查询,9,831 对分级标注,整套只花了 2.6 美元 三个结论 01|单独重排,约等于没赢 Jev 重排 bge-m3 Show more

Jason Zhu
Jason Zhu
@GoSailGlobal

有美团、阿里的老哥嘛? 试试加一路召回、重排(离线、近实时实现),我觉得有奇效 他在文本理解上 跟之前机器学习、llm很不一样 还能自动打标签做特征

Reply

Six uses that stuck after 60 days