Cortex v0.2.1 - Load files into contex, Command History and Smart spinners

✨ What's new in v0.2.1

This update is all about User Experience (UX). Cortex now feels faster, smarter, and much more comfortable for daily use.

🚀 New Features

• Load Files into Context (/load): You can now load local files directly into the chat! Use /load main.py and then ask the AI to find bugs or rewrite it. (A big step towards autonomous coding agents!).

• Command History: Press the Up and Down arrows on your keyboard to quickly navigate through your previous commands. No more retyping long prompts!

• Smart Spinners: Waiting for cloud providers (OpenAI/Gemini) is no longer boring. A smooth animated spinner now shows that Cortex is working, followed by a smart summary of the generated response.

##🛠️ Improvements

• Background Auth: Fixed a visual glitch where the token authentication message could duplicate in the terminal.

• Stability: Improved the old config migration logic to prevent 'str' object has no attribute 'get' errors when upgrading from older versions.

📖 New Commands

/load <file_path> - Attach a file to the current chat context.

📥 How to update

git pull origin main
pip install -r requirements.txt

Cortex v0.2.0 - Token auth, Advanced config, Streaming and UI toggles

🚀 Major Update: Cortex v0.2.0

This release brings massive quality-of-life improvements, a redesigned configuration system, and much better security workflow.

✨ New Features

• Seamless Authentication: Goodbye typing your password every time! Cortex now generates a secure local token on the first run. Subsequent launches are instant and password-free.
• Custom Agent Personalities: You can now rewrite the system_prompt for any agent directly in config.json. Make your code agent speak Russian, or your ideas agent act like Steve Jobs.
• UI Toggles: Added ui.show_ascii_art setting. Want a minimal, clean terminal? Just set it to false in your config.
• Context Cleanup: Added /clear command to wipe AI memory without deleting the chat.

🛠️ Architecture & Under the Hood

• Redesigned config.json: Settings are now logically grouped into providers, agents, ui, and security blocks for better readability.
• Smart Config Migration: Upgrading from v0.1.x? Don't worry. Cortex will automatically detect the old format and upgrade it to v0.2.0 on the fly.
• Chat Export: Added /export to save your encrypted chats as beautiful Markdown files.
• Live Streaming: Ollama models now print responses in real-time (character by character).
• Improved .gitignore: System files, logs, and debug chats are now strictly blocked from being committed.

📖 New/Updated Commands

• /clear - Forget the current context.
• /export - Save the active chat to a .md file.

📥 How to update

git pull origin main
pip install -r requirements.txt

Cortex v0.1.3 - Logger

🛠️ What's new in v0.1.3

Features

• Internal Logger: Added a professional logging system (logger.py). All system events, warnings and errors are now automatically saved with timestamps to ~/.cortex/cortex.log for easier debugging.

• Clean CLI: Logger output is now strictly redirected to the log file. The terminal interface remains clean and free of duplicate technical messages.

📥 How to update

git pull origin main

Cortex v0.1.2 - Live Streaming and new commands : /export and /clear

🚀 What's new in v0.2.0

• This update brings the Cortex experience to life with real-time text generation and powerful new tools to manage your AI workspace.

✨ New Features

• Live Streaming (Ollama): Local models now print responses character by character, just like ChatGPT! (Note: Cloud providers like OpenRouter use instant generation for stability).

• Export to Markdown: Use the /export command to save any encrypted chat as a beautiful .md file.
  Context Management: Added /clear command to wipe the AI's memory in the current chat without deleting the chat itself.

🛡️ Under the Hood & Fixes

• Smart Streaming Routing: Fixed a 403 Security Policy bug where OpenRouter's firewall blocked streaming requests. Cortex now intelligently uses streaming only where it works best (local Ollama).

• Full History Support for Streams: Fixed a bug where streaming mode ignored previous chat history. It now remembers the full context!

• Better Error Handling: Improved error messages for OpenRouter HTTP exceptions.

📖 New Commands

/clear - Forget the current context.

/export - Save chat to a Markdown file in your current directory.

📥 How to update

git pull origin mainpip install -r requirements.txt

Note: If you were using OpenRouter for streaming, please check your config.json - it works perfectly now without errors!

Cortex v0.1.1 - Flexfible settings agents and smart Database

🛠️ What's new in v0.1.1

✨ Features

• Flexible Agent Settings: You can now set custom temperature and max_tokens for each agent in config.json. Make the code agent strict (temp 0.2) and the ideas agent highly creative (temp 0.9)!

• Smart Database Recovery: If the encrypted database is corrupted or the password is wrong, Cortex now politely asks if you want to reset it, instead of just crashing.

🔄 Under the hood

• Added num_predict (max tokens) support for local Ollama models.

• Improved config migration (v0.1.0 configs will automatically upgrade to the new format).

Cortex v0.1.0 - Initial Release

🎉 What's new in v0.1.0

This is the first official release of Cortex!

✨ Features

• Multi-Agent System: Split tasks between code, ideas, and other agents.

• Provider Flexibility: Use local models via Ollama (tested with glm-4.7-flash) or cloud APIs (OpenAI, Gemini, OpenRouter).

• Military-Grade Security: Chat history is fully encrypted using Fernet (AES-128-CBC + HMAC). No one can read your data without the master password.

• Easy Configuration: Simple config.json setup for API keys and agent routing.

🛠️ Under the hood

• Clean Python CLI interface.

• Automated testing pipeline using pytest (3 tests for cryptography and database integrity).

• MIT License.

📥 Installation and run

git clone https://github.com/Alexx-coder/Cortex.git
cd Cortex
pip install -r requirements.txt
python runner.py