v2.0 — The "Pro" Upgrade

This is a major milestone release that transforms the tool from a simple text processor into a comprehensive AI Workstation. Version 2.0 introduces a completely redesigned UI, a dedicated Co-Writer mode, advanced Diff visualizations, and enterprise-grade local security.

✨ New Features

🧠 Co-Writer Mode

A brand new workspace dedicated to creative writing and text continuation.
Dedicated UI: A "Drawer" style interface that sits above the main editor.
Smart Presets: Built-in flows like "Universal Flow", "Narrative Velocity", and "Add Detail".
Custom Presets: Create, save, and lock your own system instructions for specific writing styles.

🎨 UI & Theming Overhaul

High-Contrast Light Mode: Completely redesigned Light theme with "GitHub-style" syntax highlighting and optimal contrast.
Deep Dark Mode: Refined color palette for long coding sessions.
Streaming FX: "Focused Beam" animation during active generation provides visual feedback on the AI's status.

⚙️ Global Settings & Localization

Multi-Currency: Track API costs in USD, EUR, or RUB (automatically synced with daily exchange rates).
Language Support: Interface fully translated into English, Russian, and German.
Model Manager: Add any OpenRouter model ID manually via the new settings UI.

🛡️ Security & Backend

AES-128 Encryption: API Keys are no longer stored in plain text. They are encrypted locally using the cryptography library (Fernet) and stored in api_key.bin.
Precise Cost Calculation: The app now reads the official usage data returned by OpenRouter API headers to calculate costs with laser precision (down to $0.000001).

🔧 Improvements & Fixes

Prompt Engineering: Updated Default.txt with a "Strict Monolingual" protocol to prevent models (like Gemini Flash) from accidentally translating text instead of editing it.

Code Block Formatting:

Improved highlight.js integration for better syntax coloring in both Light and Dark modes.
Dependencies: Switched installation method to requirements.txt for better stability.

AI Text Enhancer Pro v1.1.3 (Hotfix & Polish)

AI Text Enhancer Pro — v1.1.3

This update addresses critical interaction issues and continues refining UI clarity.

🛠 Critical Fixes

• Stop Button Logic

  • Fixed a bug where Stop remained disabled during streaming
  • Requests can now be halted instantly on click

• Correct Status Handling

  • Canceling a stream now shows a yellow “Stopped” state
    (instead of an incorrect red “Error”)

• Immediate Abort Mechanism

  • Stream termination is now instant, preventing further token output

🎨 Visual Refinements

• Generation Button State
  • While active, the button turns red (#da3633) to clearly signal Stop mode
  • Improves visibility and reduces misclicks

AI Text Enhancer Pro v1.1.2

AI Text Enhancer Pro — v1.1.2

This patch focuses on UI/UX consistency and improved workspace information clarity.

✨ New Features

Help Center (? button)
Includes:

• Project and version information
• Visual shortcuts guide with keyboard badges
• Quick links to Documentation & GitHub

Live Model Indicator

• Status Bar now displays the currently active model (accent color)
• Shows the stored Default model nearby for easier tracking of context switches

🎨 UI/UX Improvements

Full English Localization

• Toolbar labels translated for a unified, professional English interface
  (e.g., “Модель” → “Model”, “Промпт” → “Prompt”)

Status Bar Clarity

• “Default” renamed to “Def. Model” to avoid ambiguity between:
  • Active model currently running
  • Saved default configuration

Branding Consistency

• Updated About window typography and Gradient “Pro” badge for visual identity alignment

Visual Polish

• Alignment fixes for keyboard badges in Help modal
• Improved hover styles, removed dotted list markers to reduce visual noise

🛠️ Minor refinements and stability tweaks included.

v1.1.1: Smart Error Handling & UI Polish

This update focuses on stability, error visibility, and interface symmetry.

🛡️ Smart Error Handling

• Intelligent Stream Parsing: The app now intercepts API errors (e.g., 404 Model Not Found, 402 Insufficient Credits) immediately at the start of the stream.
• Clean Error Reporting: Instead of dumping raw JSON error logs into the text editor, the app now stops execution and displays a clear, human-readable alert in the status bar.
• Zero "Fake Text": Prevents the editor from treating error messages as generated content.

💎 UI/UX Improvements

• Symmetric Layout: Fixed the visual misalignment between the Input and Output panels. Both panels now share identical header structures and heights.
• New Output Toolbar: Added a dedicated toolbar for the Output pane to house actions.
• Improved Copy Action: The "Copy" button has been relocated to the new toolbar for better accessibility and cleaner aesthetics.

📝 Documentation

• Updated README.md with a revised "Key Features" section and a modular architectural description.

v1.1.0: Frontend Refactoring & Modular Architecture

🏗️ Modular Frontend Update

This release transitions the project from a monolithic file structure to a modular architecture. We have separated the design (CSS) and logic (JS) from the HTML, making the codebase cleaner and easier to maintain.

🔄 Key Changes

• Code Separation: Extracted code from index.html into dedicated files inside a new static/ directory.
  • static/styles.css: Contains all UI styling and themes.
  • static/main.js: Contains all client-side logic and API handling.
• Backend Update: Updated server.py to serve static files using absolute paths for better stability across different environments.
• Clean HTML: index.html is now lightweight and only handles the layout structure.

🛠️ Benefits

• Stability: Reduced risk of breaking scripts while editing styles (and vice versa).
• Maintainability: Easier to navigate through the project structure.
• Best Practices: Follows standard MVC-like web development patterns.

📦 How to run

python server.py

v0.2.1 Secure Core, Dynamic Models & UI Overhaul

🚀 Major Release v0.2.1

This update brings a complete overhaul to the application's security architecture and user interface. We have moved away from plain-text configuration files to a secure, encrypted local storage system.

✨ New Features

• 🔒 Secure Core: API Keys are now encrypted using AES-128 (Fernet). Keys are stored in a binary file (api_key.bin) and are never accessible in plain text on the disk.
• ⚙️ Dynamic Model Manager: You can now add, remove, and manage OpenRouter models directly from the UI. No code changes required.
• ⭐ Favorites System: Mark your favorite Model and Prompt as "Default" with a single click. They will auto-load on restart.
• ⚡ Async Backend: Completely rewritten server.py using FastAPI for better streaming performance and stability.
• 👀 Diff View: New visual comparison mode (Red/Green) to see exactly what the AI changed.
• 🛡️ Injection Protection: Enhanced XML sandboxing (<text_to_edit>) for all prompts.

⚠️ Important: Migration Guide

Since the architecture has changed, previous configuration files (openrouter_api_key.txt) are no longer supported.

1. Update dependencies:

   pip install -r requirements.txt

2. Setup API Key:

   • Launch the server: python server.py
   • Click the 🔑 Key button in the top toolbar.
   • Enter your OpenRouter key in the secure modal window.

📦 Changelog

• [Backend] Added cryptography library for key management.
• [Backend] Switched from requests to httpx for async streaming.
• [Frontend] Added Modal windows for Key and Model management.
• [Frontend] Added localStorage persistence for user preferences.
• [UI] Improved "VS Code" styling and responsiveness.

v0.2.0: Security Core & Personalization

🚀 Major Update: Security & Workflow

This release introduces a robust anti-injection security layer and a fully customizable workspace. You can now manage models directly from the UI, set defaults, and work with a secured prompt architecture.

🔥 New Features

• 🛡️ Anti-Injection Security Core:
  • Implemented an XML-tag sandboxing system (<text_to_edit>) in server.py.
  • Prevents the LLM from executing user commands (Prompt Injection) instead of processing the text.
  • Added _BOILERPLATE_UNIVERSAL.txt for creating secure custom prompts.
• ⚙️ Dynamic Model Manager:
  • Add any OpenRouter Model ID directly via the UI (no code changes required).
  • Models are saved in the browser's LocalStorage.
• ⭐ Favorites System (Defaults):
  • Set your Default Model and Default Prompt with a single click (Star button).
  • Your preferred setup loads automatically on startup.
• 📊 Enhanced Status Bar:
  • Redesigned footer for better visibility.
  • Added live indicators for active Default Model and Default Prompt.
  • Completion status now reports the model used (e.g., 🟢 Completed (GPT-4o)).

💅 UI & UX Improvements

• Added a "Clear Workspace" button with confirmation.
• Improved token counter formatting.
• Fixed UI layout for the logo and toolbar buttons.
• Added specific CSS styling for "Favorite" buttons (Golden Star).

🔧 Technical Changes

• Updated server.py to wrap user input in XML tags automatically.
• Refactored index.html to support LocalStorage persistence.

---

AI Text Enhancer Pro v0.1.0

🚀 First public release of the project.

✨ Key Features

• Live streaming responses from LLM
• Smart text enhancement (editing, polishing, rewriting)
• Code diff-view for before/after comparison
• Syntax highlighting for better readability
• FastAPI backend + clean web UI
• Supports multiple model providers (via API keys)

🔧 How to Run

pip install -r requirements.txt
python server.py

Open index.html in your browser and start enhancing text.