--- title: Build agents with Gemini API (I/O Connect ‘26) | daily.dev description: A Google DeepMind product manager and a Gemini API engineer walk through building AI agents using the Gemini API and AI Studio. Key highlights include: new AI... canonical: https://daily.dev/posts/build-agents-with-gemini-api-i-o-connect-26--lylofj4k9 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Build agents with Gemini API (I/O Connect ‘26) | daily.dev og:description: A Google DeepMind product manager and a Gemini API engineer walk through building AI agents using the Gemini API and AI Studio. Key highlights include: new AI... og:url: https://daily.dev/posts/build-agents-with-gemini-api-i-o-connect-26--lylofj4k9 og:image: https://api.daily.dev/og/posts/lYLOFJ4K9.png og:image:alt: Build agents with Gemini API (I/O Connect ‘26) og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Build agents with Gemini API (I/O Connect ‘26) **[Google for Developers](https://daily.dev/sources/googledevelopers)** · 37 min read · 1 upvotes · 0 comments ## Summary A Google DeepMind product manager and a Gemini API engineer walk through building AI agents using the Gemini API and AI Studio. Key highlights include: new AI Studio features like full-stack Node.js app generation, native Firebase integration, Google Workspace connectivity, native Android/Kotlin app generation via natural language, and one-click export to local environments. The second half covers the newly GA Interactions API — a unified interface for Gemini models and agents that introduces a steps-based data model replacing the old user/model turn concept, server-side context management, async/background execution, and support for mixing first-party tools (Google Search, Maps) with custom functions. The managed anti-gravity agent runs in isolated persistent cloud sandboxes, supports GitHub/GCS file mounting, and uses a network proxy for secure credential injection. A live demo builds an AI talk radio show agent using flat markdown skill files rather than complex Python loops. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is the Gemini interactions API and how is it different from previous Gemini API calls? The interactions API is a unified interface for using Gemini models and Gemini agents, generally available since the Monday before Google I/O Connect '26. Unlike the classic user-model turn structure, it uses a steps-based data model where each interaction type (user input, reasoning, function call, function result) is its own data type, and it supports async background execution plus optional server-side context management so you only send new inputs rather than tracking full conversation state. _Developers restructuring agent code around a new API pattern can track how it evolves via daily.dev._ ### How do anti-gravity remote agents run code without managing my own server infrastructure? Anti-gravity remote agents execute inside isolated, persistent sandboxes managed inside Google Cloud, letting the agent run commands, install dependencies, and create files without any local infrastructure. Environments can be provisioned from a GitHub repository, a GCS bucket, or inline files, including private repos and buckets, and state persists across follow-up requests so installed dependencies remain available. _Teams evaluating managed agent sandboxes can compare approaches like this one on daily.dev._ ### How does the Gemini interactions API keep agent credentials secure when agents call APIs on their own? A network proxy sits between the agent and outbound calls, letting developers restrict which URLs or domains an agent can access and inject real credentials at the proxy level via header transformation rather than exposing them inside the sandbox. This means the agent only knows it can call an API like Gemini's, never holds the actual credential, and credentials can be rotated or revoked without changing agent code. _Developers weighing agent security tradeoffs can follow credential-handling patterns like this on daily.dev._ ## Similar posts on daily.dev - [Gemini Managed Agents: Developer Guide](https://daily.dev/posts/gemini-managed-agents-developer-guide-8gahhawdu) · Philipp Schmid · 0 upvotes · 0 comments - [All the news from the Google I/O 2026 Developer keynote](https://daily.dev/posts/all-the-news-from-the-google-i-o-2026-developer-keynote-bloubqs57) · Google Developers · 4 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#google-ai-studio](https://daily.dev/tags/google-ai-studio) [View this post on daily.dev](https://daily.dev/posts/build-agents-with-gemini-api-i-o-connect-26--lylofj4k9) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Build agents with Gemini API (I/O Connect ‘26)","url":"https://daily.dev/posts/build-agents-with-gemini-api-i-o-connect-26--lylofj4k9","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/build-agents-with-gemini-api-i-o-connect-26--lylofj4k9"},"datePublished":"2026-07-06T18:42:40.886Z","dateModified":"2026-09-13T19:13:56.351Z","description":"A Google DeepMind product manager and a Gemini API engineer walk through building AI agents using the Gemini API and AI Studio. Key highlights include: new AI...","image":"https://i.ytimg.com/vi/d9LAQWKUnx8/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/d9LAQWKUnx8/sddefault.jpg","isAccessibleForFree":true,"articleSection":"Google for Developers","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Google for Developers","logo":"https://media.daily.dev/image/upload/s--jj3nxRoG--/f_auto/v1725035497/logos/googledevelopers","url":"https://daily.dev/sources/googledevelopers"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/build-agents-with-gemini-api-i-o-connect-26--lylofj4k9","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,ai-agents,google-ai-studio","timeRequired":"PT37M","video":{"@type":"VideoObject","name":"Build agents with Gemini API (I/O Connect ‘26)","description":"A Google DeepMind product manager and a Gemini API engineer walk through building AI agents using the Gemini API and AI Studio. Key highlights include: new AI...","thumbnailUrl":"https://i.ytimg.com/vi/d9LAQWKUnx8/sddefault.jpg","uploadDate":"2026-07-06T18:42:40.886Z","duration":"PT37M","url":"https://api.daily.dev/r/lYLOFJ4K9","embedUrl":"https://www.youtube.com/embed/d9LAQWKUnx8"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Google for Developers","item":"https://daily.dev/sources/googledevelopers"},{"@type":"ListItem","position":3,"name":"Build agents with Gemini API (I/O Connect ‘26)"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/build-agents-with-gemini-api-i-o-connect-26--lylofj4k9#faq","mainEntity":[{"@type":"Question","name":"What is the Gemini interactions API and how is it different from previous Gemini API calls?","acceptedAnswer":{"@type":"Answer","text":"The interactions API is a unified interface for using Gemini models and Gemini agents, generally available since the Monday before Google I/O Connect '26. Unlike the classic user-model turn structure, it uses a steps-based data model where each interaction type (user input, reasoning, function call, function result) is its own data type, and it supports async background execution plus optional server-side context management so you only send new inputs rather than tracking full conversation state. Developers restructuring agent code around a new API pattern can track how it evolves via daily.dev."}},{"@type":"Question","name":"How do anti-gravity remote agents run code without managing my own server infrastructure?","acceptedAnswer":{"@type":"Answer","text":"Anti-gravity remote agents execute inside isolated, persistent sandboxes managed inside Google Cloud, letting the agent run commands, install dependencies, and create files without any local infrastructure. Environments can be provisioned from a GitHub repository, a GCS bucket, or inline files, including private repos and buckets, and state persists across follow-up requests so installed dependencies remain available. Teams evaluating managed agent sandboxes can compare approaches like this one on daily.dev."}},{"@type":"Question","name":"How does the Gemini interactions API keep agent credentials secure when agents call APIs on their own?","acceptedAnswer":{"@type":"Answer","text":"A network proxy sits between the agent and outbound calls, letting developers restrict which URLs or domains an agent can access and inject real credentials at the proxy level via header transformation rather than exposing them inside the sandbox. This means the agent only knows it can call an API like Gemini's, never holds the actual credential, and credentials can be rotated or revoked without changing agent code. Developers weighing agent security tradeoffs can follow credential-handling patterns like this on daily.dev."}}]} ```