ollama-rich ✨

A feature-rich Ollama client with a polished terminal UI powered by Rich.

Features 🚀

• 📦 List available Ollama models in a beautiful, colored table
• 💬 Chat with models directly from your terminal
• 📝 Markdown rendering with live updates
• ⚡ Streaming support for responsive conversations
• 🔧 Simple, ergonomic CLI
• 🧰 Handy commands to pull, load, and manage models

Requirements 📋

• Python 3.7+
• Running Ollama server
• Python packages: rich, ollama, pyyaml

Installation 📦

pip install ollama-rich

From source:

git clone https://github.com/jakbin/ollama-rich.git
cd ollama-rich
pip install .

Quick start ⚡

# 1) Optional: configure defaults
ollama-rich setup -s http://localhost:11434 -m llama2

# 2) See available models
ollama-rich models

# 3) Pull a model (streaming by default)
ollama-rich pull llama2

# 4) Load a model into memory
ollama-rich load llama2

# 5) Chat
ollama-rich chat -m llama2 "Hello there!"

# Tip: short alias also available
or chat -m llama2 "Hi!" --stream

Command reference 🧭

• Global options:

  • --host Set Ollama host (default comes from config; fallback: http://localhost:11434)
  • --version Show CLI version

• Setup/config:

  • ollama-rich setup [-s --host <url>] [-m --model <name>]

• Model management:

  • ollama-rich models – list available models
  • ollama-rich model <name> – show details for a model
  • ollama-rich pull <name> [--nostream] – pull a model (stream by default)
  • ollama-rich delete <name> – delete a model (with confirmation)
  • ollama-rich load <name> – load model into memory
  • ollama-rich ps – list models currently loaded in memory

• Chat:

  • ollama-rich chat -m <name> "Your message" [--stream]
    • If -m/--model is omitted, the default model from your config is used.

You can also use the short alias or in place of ollama-rich for all commands.

Examples 🧪

# List models
ollama-rich models

# Show info for a model
ollama-rich model llama2

# Pull without streaming
ollama-rich pull llama2 --nostream

# Chat with streaming
ollama-rich chat -m llama2 "Explain transformers in 3 bullet points" --stream

Notes 🗒️

• The default host and model are stored in a YAML config under your user config directory; run ollama-rich setup to change them interactively or via flags.
• Output uses Rich for colored tables, progress bars, and markdown rendering.

License 📄

MIT