# oterm

[![Tests](https://github.com/ggozad/oterm/actions/workflows/test.yml/badge.svg)](https://github.com/ggozad/oterm/actions/workflows/test.yml)
[![codecov](https://codecov.io/gh/ggozad/oterm/graph/badge.svg)](https://codecov.io/gh/ggozad/oterm)

the terminal client for [Ollama](https://github.com/ollama/ollama), OpenAI, Anthropic, and any [pydantic-ai](https://ai.pydantic.dev/)-supported provider.

> 🚀 **oterm is now multi-provider!**
> Alongside Ollama, oterm drives OpenAI, Anthropic, Google (AI / Vertex), Groq, Mistral, Cohere, AWS Bedrock, DeepSeek, Cerebras, Grok, Hugging Face, and any OpenAI-compatible endpoint (vLLM, LM Studio, llama.cpp, OpenRouter, LiteLLM, …).
> See [What's new](#whats-new) below for the full set of changes.

## Features

* intuitive and simple terminal UI, no need to run servers, frontends, just type `oterm` in your terminal.
* supports Linux, MacOS, and Windows and most terminal emulators.
* multiple persistent chat sessions, stored together with system prompt & parameter customizations in sqlite.
* talks to Ollama, OpenAI, Anthropic, Google (AI / Vertex), Groq, Mistral, Cohere, AWS Bedrock, DeepSeek, Cerebras, Grok, Hugging Face, and any OpenAI-compatible endpoint — local (vLLM, LM Studio, llama.cpp, …) or hosted (OpenRouter, LiteLLM, …).
* tools — built-in (`shell`, `date_time`, `think`), custom Python plugins via entry points, and any [Model Context Protocol (MCP)](https://modelcontextprotocol.io) server.
* allows for easy customization of the model's system prompt and parameters.

## Quick install

```bash
uvx oterm
```
See [Installation](https://ggozad.github.io/oterm/installation) for more details.

## Documentation

[oterm Documentation](https://ggozad.github.io/oterm/)

## What's new
* **Multi-provider, via pydantic-ai (breaking).** `oterm` is no longer Ollama-only — it drives any [pydantic-ai](https://ai.pydantic.dev/)-supported provider: OpenAI, Anthropic, Google (AI / Vertex), Groq, Mistral, Cohere, AWS Bedrock, DeepSeek, Cerebras, Grok, Hugging Face, OpenAI-compatible endpoints (vLLM, LM Studio, llama.cpp, OpenRouter, LiteLLM, …), and Ollama. Set the matching API key and the provider appears in the new-chat dropdown.
* **Refreshed chat UI.** Borderless accent-driven layout, auto-growing prompt, inline `[Image #N]` attachment tokens, a collapsing thinking section, and a live token-usage footer in place of the spinner.
* **Faster streaming.** Markdown is now updated as deltas arrive instead of being re-rendered on every token, so long responses don't slow the terminal as they grow.
* **MCP rewrite (breaking).** The `mcpServers` config block adopts pydantic-ai's standard schema (compatible with Claude Desktop / Cursor). See [docs/mcp](https://ggozad.github.io/oterm/mcp/) for the full migration notes.

### Screenshots
![Splash](https://raw.githubusercontent.com/ggozad/oterm/refs/heads/main/docs/img/splash.gif)
The splash screen animation that greets users when they start oterm.

![Chat](https://raw.githubusercontent.com/ggozad/oterm/main/docs/img/chat.png)
A view of the chat interface, showcasing the conversation between the user and the model.

![Provider and model selection](https://raw.githubusercontent.com/ggozad/oterm/main/docs/img/customizations.png)
The new-chat screen, where you pick the provider and model and customize the system prompt, tools, parameters, and thinking.

![Image selection](https://raw.githubusercontent.com/ggozad/oterm/main/docs/img/image_selection.png)
The image selection interface, demonstrating how users can include images in their conversations.

![Theme](https://raw.githubusercontent.com/ggozad/oterm/main/docs/img/theme.png)
oterm supports multiple themes, allowing users to customize the appearance of the interface.

## License

This project is licensed under the [MIT License](LICENSE).