What is oterm?

oterm is a terminal client for Ollama that lets you chat with local AI models directly from your terminal. It runs on Linux, macOS, and Windows and provides a simple, intuitive UI with no need to run separate servers or front‑ends.

How to use oterm?

Open your terminal and run uvx oterm to launch the application. Detailed installation options are available in the oterm documentation.

Key features of oterm

- Start chatting instantly with oterm in any terminal.
- Supports Linux, macOS, and Windows.
- Multiple persistent chat sessions with custom system prompts and parameters, stored in SQLite.
- Full Model Context Protocol (MCP) integration – tools, prompts, and sampling.
- Works with any Ollama model, including custom models.
- Easy customization of system prompt and model parameters.

Use cases of oterm

- Chat with local LLMs entirely from the command line without additional servers.
- Use MCP tools (e.g., a git MCP server) to provide external context to the model.
- Run multiple parallel conversations, each with its own system prompt and model settings.
- Switch between different Ollama models on the fly and fine‑tune their parameters.

FAQ from oterm

How do I install oterm?

Run uvx oterm in your terminal. More installation methods are covered in the documentation.

What platforms are supported?

oterm runs on Linux, macOS, and Windows and works with most terminal emulators.

Which models can oterm use?

oterm can use any model you have pulled in Ollama, including custom models.

Does oterm support MCP?

Yes, oterm supports Model Context Protocol (MCP) tools, prompts, and sampling, along with SSE and WebSocket transports for MCP servers.

What license does oterm use?

oterm is open‑source under the MIT License.