Developers using artificial intelligence now have an easier time managing data. The latest Gemini API updates increase file size limits and expand input options. These changes aim to remove common obstacles in data handling. As a result, teams can build and scale AI applications faster.
Takeaways 🚀
✅Expanded file input options
✅Increased file size limits
✅Simplified data ingestion process
✅Enhanced scalability for developers
✅Improved AI application workflows
You can now send files to the Gemini API directly from Google Cloud Storage (GCS) or any HTTP or Signed URL. The maximum inline file size is 100MB, up from the previous 20MB limit. In this post, you’ll learn exactly what these changes mean and how they help improve the development process.
Easier Data Input for AI Models
Getting data to your model should not slow down development. In the past, large files like videos or long documents required upload to temporary storage. The Files API, for example, kept files for only 48 hours. While this worked for prototypes, it created problems for production apps that needed permanent data storage.
The new update solves this. Gemini API now lets you bring data straight from where it lives—no extra steps needed. Two new input methods support this process.
Using External URLs
The first new option lets you provide external file URLs.
- Public URLs: If you have a public PDF or image, simply paste its link in your request. The API fetches the file automatically.
- Signed URLs: For private files on Amazon S3 or Azure Blob Storage, you can create a secure, temporary signed URL. This approach keeps your data protected but also makes it easy for Gemini API to retrieve.
With these tools, you skip downloading and re-uploading files. Instead, you point Gemini API to the file and it does the rest. You save both time and server resources.
Direct Integration with Google Cloud Storage
The second improvement makes GCS integration easier. If you keep your data in Google Cloud Storage, you can now register those files directly with Gemini API.
Just give the API the GCS path for your videos, audios, or documents. The API references your storage location without needing to move actual files. This feature is especially useful when working with large datasets already organized in GCS.
Higher File Size Limit for Inline Data
Gemini API also raises the file size cap for inline data. You can now include files up to 100MB directly in your request.
This five-fold increase changes what you can do. Higher-quality images, short audio clips, or larger documents all fit within the new limit. For rapid prototyping and real-time apps, speed matters—so this update is helpful when you want a simple, fast input method.
Inline files often mean the fastest response. The new limit makes this approach possible for many more use cases.
The Benefits for Developers and AI Projects
These changes are more than technical improvements. They change how developers work.
Greater Scalability
Direct connections to cloud storage help apps grow. Production applications often handle big and growing datasets. In the past, re-uploading each file wasn’t practical. Now, you simply reference files in GCS or through secure URLs. This streamlines scaling and helps you manage expanding projects.
Better Experience for Developers
More flexible input options and fewer upload steps make day-to-day development easier. You can quickly test using inline files or build complex systems that connect to large storage buckets. The Gemini API now adapts to individual project needs. This approach saves time and reduces frustration during development.
Looking Ahead
The Gemini API’s latest updates make AI tools easier to use for everyone. Higher file size limits and support for direct storage connections remove old barriers. Now, you can work with larger files, connect to data wherever it is stored, and focus on building smarter AI solutions. These practical enhancements help you deliver projects faster and more efficiently.
