Gemini-API is a reverse-engineered asynchronous Python wrapper for the Google Gemini web application.
The tool solves the problem of programmatic access to Google Gemini's web interface by mimicking the browser-based interactions that the web app uses. It works by automatically managing authentication cookies in the background, allowing persistent connections suitable for always-on services. The wrapper exposes Gemini's capabilities through an asynchronous Python interface, supporting text generation, image generation and editing, video and audio generation, deep research workflows, system prompt customization through Gemini Gems, and integration with Gemini extensions like YouTube and Gmail. Responses are classified into distinct output types including text, thoughts, images, videos, and audio. The tool also provides streaming mode for partial output generation and includes a command-line interface for direct interaction.
Developers should adopt this tool if they need programmatic access to Gemini's capabilities without relying on an official API. It suits projects requiring image or video generation, deep research functionality, or extension-based content generation. The tool is particularly valuable for always-on services because of its background cookie refresh mechanism. Those building applications that need Gemini's multimodal generation features will find the native support for images, videos, and audio more complete than what a text-only interface would provide.
The project maintains active development with regular updates to the package distribution pipeline. The codebase follows consistent formatting standards and manages its external dependencies carefully. The tool receives ongoing attention to keep pace with changes in the underlying Gemini web application.
