Agent Customizations
Equip your agents with domain-specific plugins, skills, subagents, and hooks for instant fluency across more than 30 major life-science databases, including the 1PB AlphaGenome Atlas, UniProt, PubChem, and ChEMBL.
Explore our next generation products
Built for developers and teams in the agent-first era
Everything you need to stay up-to-date and get help
Why choose Google Antigravity for science
Google Antigravity accelerates scientific discovery with extensible agent customizations, parallel workflow execution, and grounded verification across frontier models and scientific databases.
Apply for free access
We're giving 10,000 scientists our highest usage tier in Google Antigravity and our most advanced Gemini 4 Argon model, free of charge.
Install the Science plugin in one click from the in-app Customizations marketplace.
Featured Videos
See how researchers use Google Antigravity customizations and frontier models to accelerate discovery.
What Sets Antigravity Apart
Seamless Access to DeepMind Models
Directly query Google DeepMind frontier models and databases such as AlphaGenome, AlphaFold Database, and Atlas inside your research workspace, from generating Sashimi splice-disruption plots to running structural analysis workflows.
Visualizations and Artifacts
Produce publication-ready 3D PyMOL structural renderings, reproducible analysis scripts, and verifiable scientific artifacts that are grounded in evidence and legible at every step.
Looking for the open-source customizations? Explore the repository on GitHub
Relevant Stories
“While scaling robotics pretraining runs, we hit performance drops tied to distributed data loading. Instead of manually combing through logs across multiple repositories, we used Antigravity to test several hypotheses in parallel, trace the pipeline bottleneck, and draft a fix with unit tests in an afternoon.”
“We used Antigravity to port a legacy MATLAB scientific simulation codebase into Python. Having the agent translate the numerical routines and generate comparison tests against the original MATLAB outputs made the migration much more predictable and easier to verify.”