ggozad/oterm

What it solves

It provides a dedicated terminal-based user interface (TUI) for interacting with various AI models. It eliminates the need to switch between different web interfaces or API scripts when using different model providers.

How it works

oterm acts as a client that connects to AI providers via the pydantic-ai library. It supports a wide range of providers including Ollama, OpenAI, Anthropic, Google, Groq, Mistral, and others, as well as OpenAI-compatible endpoints like vLLM and LM Studio. Users can configure API keys and select their desired provider from a dropdown menu within the chat interface.

Who it’s for

Developers and terminal-centric users who prefer a command-line environment for chatting with LLMs across multiple providers.

Highlights

• Multi-provider support: Integrates with any provider supported by pydantic-ai.
• Optimized streaming: Updates Markdown content as deltas arrive to maintain speed during long responses.
• Modern TUI: Features a borderless layout, auto-growing prompts, and a collapsing thinking section.
• MCP Integration: Supports Model Context Protocol (MCP) servers using a schema compatible with Claude Desktop and Cursor.