OpenRouter Launches create-agent-tui to Scaffold Custom Agent Interfaces in Minutes

Apr 26, 2026 · Updated May 4, 2026

OpenRouter, a unified API platform for accessing hundreds of language models, released create-agent-tui — a specialized agent skill for building custom agent harnesses. It scaffolds a project with a Terminal User Interface (TUI) and includes pre-configured logic for multi-model inference and tool calling.

This release follows a broader industry shift toward standardized agent skills that teach coding agents how to use specific developer tools. By automating the mechanical setup of an agent's loop and interface, OpenRouter lowers the barrier to building agentic applications that require switching between different frontier models for specialized tasks.

You can use the skill to customize your agent's visual style across four different templates and toggle dozens of optional features for task execution. The tool is built on the OpenRouter TypeScript SDK and is available as an open-source package in the official OpenRouter skills repository on GitHub.

View the full update on github.com

OpenRouter

@OpenRouterApr 24

Introducing "create-agent-tui" A skill for building your own agent harness + terminal UI (TUI). The skill walks you through 4 different ways of customizing the look, and supports dozens of optional features 👇 https://t.co/BVI9YUqyYe

32518

View on X

Still wondering? A few quick answers below.

OpenRouter's create-agent-tui is a specialized skill designed for AI coding agents to help developers build their own agent harnesses and terminal user interfaces. It provides a structured starting point for creating autonomous agents that can interact with users through a terminal, handling the underlying complexity of multi-model inference and tool-calling logic automatically.

The tool functions as a skill that guides a developer or an AI coding assistant through the process of scaffolding a new agent project. It is built on top of the OpenRouter Agent SDK, which provides battle-tested abstractions for managing clean code across different AI models. It walks users through four distinct ways to customize the interface's visual appearance.

Yes, create-agent-tui is available as an open-source project. The code and implementation details are hosted in the official OpenRouter skills repository on GitHub. This allows developers to inspect the underlying TypeScript code, contribute to its development, or fork the repository to create highly customized versions of their own agentic terminal interfaces and underlying logic.

Beyond basic scaffolding, the skill supports dozens of optional features to enhance an agent's capabilities. It specifically focuses on providing clean multi-model inference code, allowing the agent to switch between different language models easily. It also includes robust tool-calling logic, which is essential for agents that need to perform actions like searching the web or editing files.

This tool is primarily designed for developers and engineers who want to build personalized AI agents without writing repetitive boilerplate code. Because it is delivered as a skill for AI coding assistants, it is particularly useful for those already using tools like Claude Code or Cursor to automate their software development workflows and agentic engineering tasks.