Gemini
Label Studio Enterprise integrates with Google Gemini through the Prompts interface so you can call models like gemini-2.5-pro, gemini-2.5-flash, and gemini-2.5-flash-lite directly from your labeling projects. Once you add your Gemini API key, these models appear alongside other providers in Prompts, letting you generate annotations, iterate on prompts, and evaluate outputs against ground truth inside Label Studio.
With this connection, you can:
- Pre-label tasks using Gemini models and write results into project controls
- Use real project data to refine prompt templates before you scale u
- Compare Gemini outputs over time using run history and exports
This Gemini integration is available in Label Studio Enterprise via Prompts.
How Label Studio Connects with Gemini in Prompts
Label Studio Enterprise treats each Prompt as a reusable configuration that knows which provider to call, which Gemini model to use, and how to map input fields and outputs. When you select a Gemini model and run a Prompt:
- The Prompts UI sends your selected tasks, prompt template, and model choice to the Enterprise backend
- The backend calls the Gemini API with your key and task data
- Responses from Gemini appear in the Response panel and can be written back to your project as predictions
- You can review outputs, adjust the prompt, and then apply it at scale for assisted labeling or evaluation
