Google launches extension system for its command-line coding tool

On Wednesday, Google officially launched a new feature for its command-line AI system, Gemini CLI, allowing outside companies to integrate directly into the AI ​​product. Called Gemini CLI Extensions, the feature launches with extensions from Figma, Stripe and other companies.

The announcement comes just two days after OpenAI launched applications in ChatGPT, which also integrated third-party systems into an AI environment. But even though app access to ChatGPT is tightly curated, Gemini CLI extensions can be released without Google’s approval or involvement. Available extensions will be hosted in public repositories on GitHub and installed manually by developers.

“This open ecosystem is vital to us,” Taylor Mullen, lead engineer on the project, told TechCrunch. “Everything we do is rooted in an equitable ecosystem that everyone can participate in.”

The first extension available was for Google’s Nanobanana image generator, which was released on GitHub last week. Once installed, the extension allows users to generate images directly from the Gemini CLI terminal.

Launched in June, Gemini CLI now has more than 1 million users, according to Google, with usage heavily skewed toward software developers. Notably, Gemini CLI is widely used in the development and maintenance of its own codebase, closely supervised by product managers, as detailed in a recent TechCrunch interview.

In an interview, Ryan J. Salva, senior director of product management for developer tools at Google, told TechCrunch that the goal of the new feature was to turn Gemini CLI into “a platform for extensibility, a conduit to other tools and instructions from elsewhere in your toolchain.”