convez.ai
ComandoUtilidadesIntermediário
DefamationStation

RetroChat v2

DefamationStation/Retrochat-v2

CLI unificada para múltiplos provedores de IA com RAG e edição de histórico

Ver no GitHub
InicialPluginScore · de 0 a 100

Interface de linha de comando para interagir com múltiplos modelos de IA (Ollama, Claude, OpenAI, Google, OpenRouter). Oferece gerenciamento de conversas, sistema RAG para documentos locais, edição de histórico e troca dinâmica de provedores.

Instalação

$npx convez-ai DefamationStation/Retrochat-v2

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
88
Forks
5
Acompanhando
4
Saúde
28%
Último commit
há 1 ano
Versão
—
Issues abertas
1
Contribuidores
2

Casos de uso

  • Experimento com diferentes modelos de IA sem trocar de aplicação
  • Processamento e consulta de documentos locais via RAG
  • Desenvolvimento e prototipagem rápida com chat multimodelo

Veredito

Ferramenta prática para quem trabalha com múltiplos provedores de IA. A integração com RAG é útil, mas depende de Ollama e nomic-embed-text especificamente. Código 100% gerado por IA pode apresentar limitações em edge cases.

README

A tradução para o português deste README ainda está na fila.

RetroChat v2

RetroChat Screenshot RetroChat is a powerful command-line interface for interacting with various AI language models. It provides a seamless experience for engaging with different chat providers while offering robust features for managing and customizing your conversations.

🚀 Installation

One installer for all platforms (Windows/Linux/macOS):

Windows (PowerShell)

powershell -ExecutionPolicy Bypass -Command "iwr -useb https://raw.githubusercontent.com/DefamationStation/Retrochat-v2/main/install.ps1 | iex"

Linux/macOS

curl -fsSL https://raw.githubusercontent.com/DefamationStation/Retrochat-v2/main/install.ps1 | pwsh -

Then run rchat from anywhere! 🎉

What the installer does:

  • Downloads to ~/.retrochat/
  • Sets up Python environment
  • Installs dependencies
  • Creates global rchat command
  • Enables auto-updates

Interactive installation:

# Download first, then run with options
./install.ps1 -Interactive
./install.ps1 -Force
./install.ps1 -CustomPath "C:\MyApps\RetroChat"

✨ Features

  • 🤖 Multi-Provider Support: Ollama, Anthropic, OpenAI, Google, OpenRouter
  • 📝 Document RAG: Load and query local documents (PDF, Word, Markdown, TXT)
  • 🔄 Auto-Updates: Always get the latest features automatically
  • ⚙️ Customizable Settings: Fine-tune AI behavior with adjustable parameters
  • ✏️ Conversation Editing: Edit entire conversation history
  • 📋 Multi-line Input: Enter complex queries with ease
  • 🎯 Command System: Control everything with '/' prefixed commands
  • 🔢 Token Counting: Track usage for messages and conversations
  • 🔀 Provider Switching: Switch between AI providers mid-conversation

📚 Document RAG System

ATTENTION: RAG only works with nomic-embed-text on Ollama currently.

  1. Create a folder in ~/.retrochat/ and put your files in it
  2. Type /load <folder_name> to process documents
  3. Use @<folder_name> <question> to query documents

🛠️ Manual Installation

For development or custom setup:

  1. Clone the repository: git clone https://github.com/DefamationStation/Retrochat-v2.git
  2. Create virtual environment: python -m venv venv
  3. Activate environment: venv\Scripts\activate (Windows) or source venv/bin/activate (Linux/macOS)
  4. Install dependencies: pip install -r requirements.txt
  5. Run: python retrochat.py
    cd C:\Users\<your username>\.retrochat
    
  6. Install the required packages:
    pip install -r requirements.txt
    
  7. Run the script with the setup flag to configure RetroChat:
    python retrochat.py --setup
    

NOTE: If chromadb doesn't properly install for you, all you need to do is download and install Microsoft Visual Studio Build Tools and after installing that navigate to the individual components and install MSVC v143 - VS 2022 C++ x64/x86 build tools (Latest) and Windows 11 SDK (10.0.22621.0)

Usage

After installation, you can start RetroChat by running the shortcut command:

rchat

This command can be used from any directory in your terminal.

Commands

RetroChat supports various commands to manage your chat sessions and settings:

  • /chat rename <new_name> - Rename the current chat
  • /chat delete - Delete the current chat
  • /chat new <chat_name> - Create a new chat
  • /chat reset - Reset the current chat history
  • /chat list - List all available chats
  • /chat open <chat_name> - Open a specific chat
  • /set system <message> - Set the system message
  • /set - Show available parameters and their current values
  • /set <parameter> <value> - Set a parameter
  • /edit - Edit the entire conversation
  • /show length - Display the total conversation tokens
  • /show context - Display the context of the last query
  • /switch - Switch to a different provider or model
  • /help - Display the help message
  • /exit - Exit the program
  • /load <folder_name> - Load documents from a specified folder
  • @<folder_name> <query> - Query loaded documents from a specific folder

Document Loading and Querying

RetroChat now supports loading and querying documents:

  • Use /load <folder_name> to load documents from a specific folder in your .retrochat directory.
  • Query loaded documents using @<folder_name> <your question>.
  • Supported file types include .txt, .pdf, .doc, .docx, and .md.

Configuration

RetroChat uses a .env file to store configuration settings. This file is automatically created in the .retrochat directory in your home folder. You can manually edit this file to set API keys and other preferences.

Updates

RetroChat checks for updates automatically when you start the application. If updates are available, you'll be prompted to install them.

Watch the demo video

Contributing

Contributions to RetroChat are welcome! Please feel free to submit pull requests or create issues for bugs and feature requests.

License

MIT License.

Ver o original no GitHub

Bom para

Relacionados