ollama-agentic 0.1.0

A beautiful, agentic CLI for Ollama — run local LLMs with auto tool-calling, memory, and more

Ollama CLI — Terminal UI for Local LLMs

A beautiful, agentic CLI for running local Ollama models with tool-calling, memory, planning, and compare modes.

Quick links

• src/main.py — main v3 implementation (contains OllamaCLI and main() entry)
• ollama_cli.py — simpler v1 CLI
• ollama_cli_v2.py — v2 CLI with themes and expanded features
• pyproject.toml — project metadata & dependencies

Requirements

• Python >= 3.10
• Install runtime deps listed in pyproject.toml: pip install rich prompt_toolkit ollama requests beautifulsoup4

Run

• From the package entry (if installed): ollama-cli (configured in pyproject.toml: ollama_cli.main:main)
• Directly:
  • v3: python src/main.py
  • v2: python ollama_cli_v2.py
  • v1: python ollama_cli.py

Features

• Interactive chat with streaming output and Markdown rendering
• Agentic mode: auto tool-calling (shell, file, fetch, ls) and iterative plan execution
• /plan: generate and execute step plans
• /run: debug loop for running and auto-fixing Python scripts
• Long-term memories (/remember, /memories, /forget)
• Model management and side-by-side model comparison
• Save/load conversations and personas

Key symbols

• OllamaCLI — primary CLI class and command implementations
• bootstrap — ensures Ollama is installed/running and can pull models
• _run_tool — executes shell/file/fetch/ls tool calls

Config & data

• Config stored at: ~/.ollama_cli_config.json
• History: ~/.ollama_cli_history
• Saves: ~/.ollama_cli_saves
• Personas: ~/.ollama_cli_personas
• Memory: ~/.ollama_cli_memory.json

Development

• Run unit tests (if any) via pytest (dev deps in pyproject)
• Linting and formatting recommended with tools of choice

Notes

• The CLI requires the ollama binary or library; v3 bootstraps installation attempts via bootstrap in src/main.py.
• Web fetch features require requests and beautifulsoup4 (optional).

License

• MIT — see LICENSE

Contributing

• PRs and issues welcome. Keep changes focused and include tests where appropriate.