$ npx @ppwnr88/noyap init

Noyap

Less yap. More code.

Multilingual communication rules for AI coding agents.

AI coding agents talk too much. Noyap makes them concise, code-first, multilingual, and safer for real developer workflows.

Get startedGitHubnpm
iterm - noyap init

➜ npx @ppwnr88/noyap init

✔ Generated AGENTS.md

✔ Thai-native mode enabled

✔ Safe merge strategy enabled

before

Certainly! The issue appears to be that your object is recreated...

after

Object recreated every render. Use useMemo.

install

CLI-first setup

shell
npx @ppwnr88/noyap init

before / after

Shorter without losing the fix

English

before

Certainly! The issue appears to be that your object is recreated on every render, which may cause unnecessary rerenders.

after

Object recreated every render. Use useMemo.

Thai

before

ปัญหานี้อาจเกิดจาก cache ของ CDN ยังไม่ถูก purge หลัง deploy

after

น่าจะติด cache หลัง deploy. ลอง purge CDN.

Mixed Thai/English

before

sequelize include ดึง latest row ยังไงดีครับ

after

include ไม่ guarantee latest row. ใช้ separate + limit/order.

agents

Works with the coding agents developers already use

$Codex
$Claude Code
$Cursor
$OpenCode
$Windsurf
$GitHub Copilot
$Cline
$Continue
$Gemini CLI
$Roo Code

thai-native

Not translation. Natural developer Thai.

language detection
Noyap detects English, Thai, and mixed Thai/English prompts, then keeps the answer in the language style that feels natural for the developer.
thai dev phrasing
It preserves common Thai developer phrasing instead of forcing formal translated prose.
deploy ยังไงcache ค้างbuild failquery ช้าroute ชนtoken หมด

safe merge

Respects existing AGENTS.md rules

AGENTS.md strategy
Noyap supports merge, separate, overwrite, and cancel. Existing project instructions stay authoritative, especially warnings and team rules.
codex
npx @ppwnr88/noyap init --agent codex --agents-md-strategy merge

modes + presets

Tune strictness by workflow

modes
minimalbalancedseniorthai-devbilingualhardcorehardcore-th
presets
backendfrontenddevopssqlsecurityreviewer

benchmarks

Measured, linked, and conservative

Token reduction

54.12%

balanced mode, benchmark suite

Character reduction

58.23%

balanced mode, benchmark suite

Line reduction

66.67%

fixed benchmark fixtures

Warning preservation

100%

heuristic checks; review critical output

Benchmark data comes from the repository fixture suite. These are not guaranteed savings for every project. See benchmark report.

ready

noyap init

Install concise agent rules in one command.

Start with safe defaults, then tune by agent, language, mode, or preset.

shell
npx @ppwnr88/noyap init