paulrobello/parllama

What it solves

ParLlama is a Terminal User Interface (TUI) application that simplifies the management and use of Large Language Models (LLMs). It provides a unified interface for interacting with local models via Ollama and various cloud-based providers, removing the need to switch between different CLI tools or web interfaces.

How it works

Built with Textual and Rich, the application acts as a client that connects to LLM providers. It integrates natively with Ollama for local model pulling, deletion, and quantization, while also supporting cloud APIs like OpenAI, Anthropic, and Groq. It features a session management system to save and organize conversations, a memory system for persistent user context, and a secure vault for API keys.

Who it’s for

Users who prefer working in the terminal but want a rich, visual interface for managing local and remote LLMs, as well as developers who need a centralized tool for chatting with multiple AI providers.

Highlights

• Multi-Provider Support: Works with Ollama, OpenAI, Anthropic, Groq, XAI, OpenRouter, Deepseek, and LiteLLM.
• Vision Capabilities: Supports chatting with images using vision-capable models like LLaVA.
• Local Model Management: Pull, delete, and quantize models directly from the TUI.
• Session & Memory: Save/load chat sessions and maintain persistent user context across conversations.
• Fabric Integration: Ability to import and use Fabric patterns as reusable prompts.
• Secure Secrets Vault: Encrypted storage for sensitive API keys using AES-GCM.
• Template Execution: A secure system for executing code with configurable allowlists.