oterm is a Python-based terminal client for multiple LLM providers (Ollama, OpenAI, Anthropic, etc.) using pydantic-ai. It supports multi-provider chat, refreshed UI, and faster streaming with delta updates.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
oterm is the terminal client for Ollama, OpenAI, Anthropic, and any pydantic-ai-supported provider. It supports multi-provider setup and provides a refreshed chat UI with auto-growing prompts and a live token-usage footer. The project is written in Python and licensed under MIT.
How it works
The tool connects to various LLM providers via pydantic-ai integration, enabling chats across multiple backends. It includes a new-chat dropdown that shows the provider once keys are configured. It streams tokens with deltas and updates markdown progressively.
Getting started
Install:
uvx oterm
Full install methods, configuration, and usage are available in the Documentation: https://ggozad.github.io/oterm/
Recent releases
- 0.23.0 (2026-07-31): Remember the provider used for the last created chat; Add a keybinding for copying the last message.
- 0.22.0 (2026-07-31): Qualify MCP tool names by server.
- 0.21.0 (2026-07-25): Selectable pydantic-ai capabilities; Label attached images with their sniffed media type.
- 0.20.0 (2026-06-30): Update to pydantic 2.x; Pre-populate system prompt and parameters from Modelf.
- 0.19.0 (2026-06-08): Forward thinking toggle to Ollama models; Bind Ctrl+t for session-only thinking toggle.
Traction
Stars: 2413; Forks: 138; Open issues: 6. (No further trend data provided in this entry.)
Behind the repo
Not provided in the supplied data.
Caveats
License: MIT. Created: 2023-10-10. Last push: 2026-07-31. Language: Python.