What categories of experiments does Google Labs cover?
Google Labs groups its experiments into four main categories: Create, Develop, Explore, and Learn. The site's navigation lists these as the top-level filters ("All Create Develop Explore Learn"), and individual experiments are described in ways that map onto them — creative studios, developer tooling, research instruments, and learning tools. The categories are a browsing aid rather than a strict taxonomy, so a few experiments sit at the edges (for example, an AI design canvas is both a creative and a developer-adjacent tool).
Create: generative and design tools
This category covers experiments for producing content — stories, music, images, UI, and marketing material.
| Experiment | What it does |
|---|---|
| Google Flow | AI creative studio for bringing stories to life |
| Google Flow Music | AI music studio for creating full songs |
| Dreambeans | Personalized daily collections of stories |
| Stitch | Turns natural language into high-fidelity UI on an AI design canvas |
| Pomelli | Builds scalable, on-brand marketing content |
| Mixboard | Concepting board to explore, expand, and refine ideas |
Develop: tools for building and shipping
These experiments target people writing code, building apps, or evaluating models.
- Jules — an asynchronous coding agent that automates tasks such as bug fixes, writing tests, and adding features.
- Stax — an AI evaluation toolkit for testing models and prompts before shipping.
- Opal — builds, edits, and shares AI mini-apps using natural language.
- Play with Putty — a research experiment in collaborative "vibe coding" where teams build tools and websites together in real time (join via waitlist).
Explore: research and discovery instruments
This group is aimed at research workflows — finding literature, generating hypotheses, and accelerating experimentation.
- Literature Insights — finds papers, structures data tables, and creates artifacts like reports and slide decks.
- Hypothesis Generation — a multi-agent tool that simulates the scientific method to identify knowledge gaps and propose novel hypotheses.
- Computational Discovery — an agentic research engine that generates and scores code variations to help discover models and speed up iteration.
Learn: education and skill development
- Learn Your Way — transforms content into a dynamic, tailored learning experience.
- Vantage — develops and measures skills like collaboration, creativity, and critical thinking through GenAI-simulated teamwork.
- AI Edge Eloquent — turns raw, unpolished speech into well-crafted written text.
How to use the categories
On labs.google, the category labels act as filters over the full experiment list. If you know what you want to do — make something, build something, research something, or learn something — pick the matching category and browse from there. Each experiment card links to either a "Try It Now" action or a "Learn More" page, and some (like Play with Putty) route through a waitlist instead of direct access. Availability and access conditions are set per experiment, so check the individual card rather than assuming every listing is immediately usable.