Imagent is an innovative tool designed to empower AI agents with the capability to create images, videos, and speech seamlessly within their workflows. It presents a unified interface that abstracts the differences between various providers and models while ensuring an organized approach to asset management for easy reuse. This functionality stands in stark contrast to conventional practices where generated content is often lost after use.

Key Features

• Generation as an Agent Capability: Imagent enables compatible agents to utilize the imagent Command-Line Interface (CLI) to generate images, videos, and speech as integral steps in their workflows. This eliminates the need for custom integration, allowing for efficient and straightforward operations.

• Unified Provider and Model Interface: With support for leading platforms such as OpenAI, Azure OpenAI, Google Imagen/Gemini, and others, Imagent offers a consistent interface where users can switch providers or models without the hassle of rewriting prompts or adjusting commands.

• Durable Asset Library: Every generated output—including images, videos, characters, and backgrounds—is stored in a local library. This allows users to curate, search, and reuse assets across various projects, enhancing productivity and creative continuity.

Getting Started

The CLI can be installed with the following command:

npm install -g @imagent/cli
imagent doctor

For the desktop application, installers for macOS and Windows are available from the latest release.

Example Usage

Generate content effortlessly with default settings:

imagent image generate "minimal product photo of a ceramic mug"
imagent video generate "a slow dolly shot through a rainy alley"
imagent speech synthesize "Welcome to imagent, your local creative workspace."

Integration with AI Agents

Imagent includes an installable skill located at skills/imagent that can be integrated into any compatible agent runtime. This streamlines the process for agents like Claude Code, Codex, and Hermes, allowing them to access local galleries and predefined provider configurations.

Typical Workflows

• Enhance coding or automation agents to generate multimedia assets mid-process using a single, recorded CLI.
• Switch model or provider for identical prompts without modifying calling methods.
• Create a library of reusable creative assets that grow through collaborations on multiple projects.
• Maintain a complete history of generated content rather than losing it after execution.
• Combine terminal automation and desktop management through a shared local workspace.

Project Structure

The repository structure includes:

imagent/
  apps/
    desktop/      # Electron desktop application
    cli/          # Command-line interface
  packages/
    core/         # Core logic
    providers/    # Provider adapters
    persistence/  # Data management
    config/       # Configuration handling
    ipc/          # Inter-process communication
    ui/           # Shared UI components

Current Status

Imagent is in early development, with evolving features and data structures. At this stage, it does not support telemetry, automatic updates, or cloud synchronization. Note that desktop applications are unsigned; users may need to adjust security settings upon first launch.

Additional Resources

• Documentation for comprehensive guidance and troubleshooting.
• Explore the options of the Desktop app and CLI.
• Check the Architecture for insights into the system design.