Forged in the span of 1 Days - Developed
Yapper
YAPPER is a privacy-first, client-side web application designed to bridge informal, colloquial Indonesian communication and high-performance LLM engineering. It transforms raw, unstructured, or slang-heavy Indonesian thoughts into dense, telegraphic English system prompts and fluent, idiomatic translations, minimizing prompt token consumption while maximizing prompt adherence and instruction fidelity across frontier models.
WebsiteLLMs
Click to reveal
The YAPPER interface presents a focused, split-pane developer workspace centered around a dual-mode toggle (Prompt Optimization vs. Natural Translation). The input panel on the left
captures raw Indonesian or conversational text with live word, character, and token estimation counters alongside quick-start example pills, while the center action bar initiates
transformation via hotkey (⌘↵). The output panel on the right renders the real-time streamed English result accompanied by a live token compression metric, one-click clipboard copy, and
character telemetry. Anchoring the top navigation are a real-time LLM provider connectivity badge, a slide-out History Drawer (⌘H) for searching and reloading past conversions, and a
Settings Modal to configure API keys, temperature parameters, and custom endpoints across providers like Groq, OpenAI, Gemini, and OpenRouter.
Built With
HTML/Javascript/CSS