• https://t.me/Lyra_IVbot

Lyra — AI Prompt Optimizer Telegram Bot

Transform rough ideas into precision-crafted prompts. Built on Cloudflare Workers.

Lyra is a Telegram bot that takes raw user input and turns it into optimized, professional-grade prompts for any AI platform. It uses a 4-D methodology (Deconstruct → Diagnose → Develop → Deliver) to analyze intent, fill gaps, and produce structured prompts with copy-friendly code blocks.

---

Features

• ✍️ Free-Form Mode — Send any text, get a fully optimized prompt instantly.
• 💻 Category Presets — Code, Image, Video with specialized system prompts and canned presets.
• 🎯 Entity-Based Formatting — Bold, italic, code blocks, and expandable blockquotes via MessageEntity arrays (no Markdown/HTML parsing).
• 🌐 Multilingual — English, فارسی (Persian), Русский (Russian) with automatic fallback.
• ⚡ Blazing Fast — Runs on Cloudflare Workers with in-memory caching and KV persistence.
• 🔄 Smart Follow-Ups — AI generates targeted questions; subsequent messages refine the prompt.
• 📋 One-Tap Copy — Native copy_text button for instant prompt extraction.

---

Architecture

┌──────────┐     ┌──────────────────┐     ┌─────────────┐
│ Telegram │────▶│  Cloudflare      │────▶│  OpenRouter │
│ Bot API  │◀────│  Worker (index)  │◀────│  AI APIs    │
└──────────┘     └──────────────────┘     └─────────────┘
                      │
                      ▼
               ┌──────────────┐
               │  KV Storage  │
               │  (state)     │
               └──────────────┘

Layer	File	Responsibility
Entry	src/index.js	Webhook router, debug endpoints, env check
Bot	src/bot.js	Message routing, state machine, response assembly
API	src/telegram.js	HTTP client, entity builders, message helpers
i18n	src/messages.js	Multilingual strings with function-based interpolation
UI	src/keyboards.js	Reply keyboards, inline keyboards, copy button
AI	src/openrouter.js	Model fallback chain, API call with timeout
Prompts	src/templates.js	4-D base system prompt, category presets

---

Project Structure

src/
├── index.js          # Worker entry, webhook handler, debug endpoints
├── bot.js            # Message routing, state machine, command handling
├── telegram.js       # Telegram API client, entity formatting helpers
├── messages.js       # i18n strings (en/fa/ru)
├── keyboards.js      # Reply & inline keyboard builders
├── templates.js      # AI system prompts (BASE_4D, categories, presets)
└── openrouter.js     # OpenRouter API with model fallback chain

---

Quick Start

Prerequisites

• Node.js 18+
• Wrangler CLI (npm install -g wrangler)
• A Telegram Bot Token
• An OpenRouter API Key

1. Clone & Install

git clone <repo-url> lyra-bot
cd lyra-bot
npm install

2. Configure Environment

Set these secrets in your Cloudflare Worker or .dev.vars:

Variable	Description
TELEGRAM_TOKEN	Telegram Bot API token (from BotFather)
OPENROUTER_API_KEY	OpenRouter API key
OPENROUTER_MODEL	(optional) Preferred model ID
BOT_LANGUAGE	(optional) Default language (en, fa, ru)
DEBUG_KEY	(optional) Auth key for /debug/* endpoints
LYRA_STATE	(bound KV namespace) Persistent state storage

3. Run Locally

npx wrangler dev

The dev server starts at http://localhost:8787. Use /setup to register the webhook:

GET http://localhost:8787/setup

Or manually:

curl -X POST https://api.telegram.org/bot<TOKEN>/setWebhook \
  -d "url=https://your-worker.workers.dev/webhook"

4. Deploy

npx wrangler deploy

---

API Endpoints

Method	Path	Description
POST	/webhook	Telegram update webhook (main entry)
GET	/ or /status	Health check
GET	/debug/overview	System status overview
GET	/debug/telegram	Bot info & webhook status
GET	/debug/openrouter	OpenRouter API key validation
GET	/debug/kv	KV namespace entries
GET	/debug/env	Environment variables (sanitized)
GET	/debug/test-update?chat_id=...&text=...	Test message delivery

Access debug endpoints with ?key=<DEBUG_KEY> (if DEBUG_KEY is configured).

---

Usage

/start — Main Menu

🚀 Lyra — AI Prompt Optimizer

I transform your rough ideas into precision-crafted prompts.

Just send any text, or pick a mode below ✨

Inline keyboard: Quick Prompt, Code, Image, Video, Guide, Language.

Free-Form Mode

Simply type any text. Lyra applies the BASE_4D system prompt and returns:

📥 Your Input
> your original text

📤 Optimized Prompt
```code block```

Lyra — Crafted with Precision

Category Mode

Tap Code, Image, or Video to choose preset system prompts tailored to that domain.

Follow-Up Refinement

When the AI generates follow-up questions, just continue typing to refine the prompt. The bot merges your additional context and re-optimizes.

---

Message Formatting

Lyra uses MessageEntity arrays exclusively — no parse_mode (MarkdownV2 or HTML). This eliminates escaping issues and gives precise control over formatting.

Entity	Usage	Helper
bold	Labels, emphasis	B(text)
italic	User input, notes	I(text)
code	Inline snippets	C(text)
Pre-formatted	Prompt output	buildPreBlock(text)
Expandable blockquote	Footer, notes	buildBlockQuote(text)

---

AI System Prompt

The BASE_4D template (src/templates.js) implements Lyra's core methodology:

1. Deconstruct — Extract intent, entities, context
2. Diagnose — Audit clarity, specificity, completeness
3. Develop — Apply optimal techniques per request type
4. Deliver — Construct formatted prompt inside <PROMPT> tags

The AI outputs wrapped in <PROMPT>...</PROMPT> and optional <FOLLOWUP>...</FOLLOWUP> tags, which the bot parses and presents cleanly.

---

Language Support

Messages are defined in src/messages.js as either plain strings or functions (for interpolated values). The getMsg(lang, key, ...args) helper:

1. Looks up MESSAGES[lang][key]
2. Falls back to en, then ru
3. Returns the key name if nothing is found

---

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/my-idea)
3. Commit your changes (git commit -m 'Add my idea')
4. Push to the branch (git push origin feature/my-idea)
5. Open a Pull Request

---

License

MIT