Agent Launcher

What Agent Launcher does

Agent Launcher supports Claude Code, Codex CLI, OpenCode, Pi, Gemini CLI and Hermes Agent. On first launch a wizard checks the local environment, links existing binaries and walks through official-account or API configuration. A CLI it cannot find can be installed in one click, with Hermes Agent using its official installer, while a CLI you already have is never reinstalled or updated. For Hermes the configuration options are an existing account or OpenAI-compatible providers.

Each agent can hold several provider profiles. When a profile changes, the app synchronizes the environment variables and native config files that CLI expects, including Hermes Agent config and env files, and masks secrets in previews. A minimal real model request tests the endpoint, credentials, model and account before a profile is used. Users pick a project folder before a session starts, and the app reads each CLI's own local conversation history so sessions can be resumed or deleted. Installed MCP servers and Skills can also be inspected.

Key features

• Detects and links CLIs already installed on the system
• One-click install of missing CLIs, using Hermes Agent's official installer
• Multiple provider profiles per agent with a real connectivity check
• Syncs environment variables and native config files, with secrets masked in previews
• Embedded terminal or chat view with project-aware sessions
• Reads each CLI's local session history to resume or delete sessions

When to use it

• Switching Hermes Agent between provider profiles without editing config files by hand
• Running several coding agents against the same project folder
• Checking that an API endpoint and model work before using a profile

Who it is for: Developers who use several coding-agent CLIs and prefer to manage them from one desktop app.

How it fits with Hermes Agent

Hermes Agent is one of the six agents the launcher supports, using an existing system installation and writing its config and env files.

FAQ