Google Makes the Interactions API the Default Way to Use Gemini Models and AI Agents
Google has a new main way to talk to its Gemini AI models. It is called the Interactions API. An API is a set of rules that lets one app talk to another app. Coders use an API to plug their apps into an AI model. So this is now the main door into Gemini. Coders use it to build apps and AI agents. Google shared this news on June 22, 2026.
This new tool was in beta since December 2025. Beta means it was an early test version. Only some people could try it then. Now it is “generally available.” That means anyone can use it.
What is the Interactions API?
The Interactions API is Google’s way to use its Gemini models and AI agents. An AI agent is software that can do tasks on its own. It does not just answer one question. It can take many steps to finish a job.
It takes the place of an older tool called generateContent. The old tool still works for now. But Google is slowly turning it off. From now on, all new agent features will come only through the Interactions API.
Google also moved its own coder tools to the new API. It is now the default in Google AI Studio. It is the default in all of Google’s help pages too.
What changes for developers?
The biggest change is how the API is built. A developer is a person who writes code to build apps. The old way used simple labels like “user” and “model.” These labels showed who said what.
The new way uses clear “steps” instead. Each action becomes its own step. A user typing something is one step. The AI calling a tool is another step.
This step-by-step design fits AI agents well. That is because agents take many actions in a row to get a job done.
Two cost-and-speed modes
The new API has two modes. Flex mode cuts the cost by 50 percent. So it is the cheaper choice. Priority mode is built for speed. So it answers faster. Developers can pick the mode that fits their app.
Key facts
| Item | Detail (as reported) |
|---|---|
| Beta launch | December 2025 |
| General availability | June 22, 2026 |
| Replaces | The generateContent interface |
| Flex mode | Cuts costs by 50% |
| Priority mode | Optimized for speed |
| Default in | Google AI Studio and all docs |
New features built in
The Interactions API comes with new tools to help build agents:
- Managed Agents that run inside a Linux sandbox. A sandbox is a safe, walled-off space. Code can run there without harming the main system.
- Background execution. This means long jobs can keep running while the user does other things.
- Tool chaining. This links Google Search and Google Maps into one smooth flow.
- Media generation. This makes images, music, and speech.
Google’s developer relations lead spoke about the change. This is the person who helps coders use Google’s tools. They said the new tool “sets the stage for the new era of Agents.” Google also put out a migration guide. A migration guide is a step-by-step help page. It shows coders how to move from the old API to the new one.
Why it matters (especially for India and founders)
India has a huge group of coders. It also has many fast-growing AI startups. A startup is a new, small company. Many of them build their apps on top of Gemini. So this change touches them in a direct way.
Flex mode cuts costs by 50 percent. This can help small teams save money. A founder is the person who starts a company. For a founder who watches every rupee, smaller API bills help a lot. Lower bills make it easier to test new ideas. The built-in sandbox and background tasks help too. They mean less work building these tools from scratch.
There is one catch. It is the migration. Migration means moving from the old tool to the new one. Teams that use the old generateContent path will need to plan this move. Doing it early is smart. It helps you avoid a last-minute rush when the old tool is finally shut off.
FAQ
Is the old generateContent interface gone?
No. It still works. But Google is turning it off slowly. New agent features will come only through the Interactions API.
What is the main benefit of the new API?
It is built for AI agents. It uses clear “steps” for each action. It also adds tools like sandboxes, background tasks, and media generation.
Can it save money?
Yes. Flex mode is said to cut costs by 50 percent. That is compared with the standard option.
Do I need to migrate right now?
Not right away. The old tool still works. But Google has put out a migration guide. Moving early is the safer path.
Takeaway
Google is moving all Gemini work onto one clear path. This path is built for AI agents. The Interactions API is now the default. It is cheaper in Flex mode. And it comes with the agent tools coders will need next. Teams should read the migration guide. They should start planning their move now.
Source: The Decoder
