The Antigravity SDK (Python) is Google’s Python library for building AI agents on the Antigravity platform and Gemini models. Note that this is a distinct project from the Google Antigravity product already listed separately on Tooloso: this entry covers the open-source Python SDK repository, not the Antigravity application itself.
The SDK is developed in the open on GitHub, where it has about 3,810 stars, with active development in 2026. It provides a secure, scalable and stateful infrastructure layer that abstracts the agentic loop, so developers can focus on what an agent should do rather than how it runs. The library ships as platform-specific wheels on PyPI that include a compiled runtime binary, and it integrates with MCP tools and reusable agent skills. It can run against the Gemini API with a simple key, or against Gemini Enterprise Agent Platform (formerly Vertex AI) using express mode or Application Default Credentials. Example projects in the repository, including a hello-world agent, show the intended workflow end to end, from authentication through tool use, which shortens the path from installation to a working agent connected to real services.
The SDK itself is free and open source under the Apache-2.0 license. Usage costs come from the underlying Gemini API or Google Cloud services your agents call.
Python developers who want to build production agents on Google’s Antigravity runtime and Gemini models without hand-rolling the agent loop, state management and tool wiring themselves.