This tutorial shows you how to publish AI agents to Google Workspace as Google Workspace add-ons, using Apps Script or HTTP endpoints. After your publish your add-on, your users can interact with the AI agents within their workflows.
Google has released a comprehensive new tutorial that demonstrates how to bridge the gap between complex AI agents and the daily workflows of Google Workspace users. The Travel Concierge sample uses the Agent Development Kit (ADK) to deploy a conversational, multi-agent AI directly into the Google Workspace platform.
While many AI demos focus on standalone chat interfaces, this solution stands out by implementing the agent as a Google Workspace Add-on. This means the AI isn’t just a browser tab; it is a helpful sidebar accessible directly within Gmail, Calendar, Drive, Docs, Sheets, and Slides, as well as a fully functioning Google Chat app.
Key Features for Developers
The sample provides a robust blueprint for developers looking to build “agentic” workflows. Key capabilities include:
Multi-Agent Architecture: The solution uses Vertex AI to manage a “Travel Concierge” that orchestrates sub-agents for specific tasks, utilising tools like the Google Maps Platform Places API and Google Search Grounding.
Context Awareness: In applications like Gmail, the agent can read the context of the currently selected email to help plan trips based on actual correspondence.
Rich User Interfaces: The add-on moves beyond simple text, utilising the Card framework to display rich widgets and interactive elements.
Cross-Platform Persistence: User sessions are managed by Vertex AI, ensuring that a conversation started in Google Chat can be seamlessly continued in a Google Doc sidebar.
Flexible Implementation
The Apps Script implementation demonstrates how to connect a standard Workspace Add-on to the powerful Vertex AI backend using UrlFetchApp. It serves as an excellent reference for handling synchronous execution limits and rendering complex agent responses within the constraints of the Add-on sidebar.
Whether you are looking to build a travel planner or a complex enterprise assistant, this sample provides the architectural patterns needed to bring your AI agents to where your users actually work.
A fact-check custom function for Google Sheets to be used as a bound Apps Script project powered by a Vertex AI agent and Gemini model.
This sample demonstrates how you can use two powerful types of AI resources directly into your Google Sheets spreadsheets
A new tutorial published by Google’s Pierrick Voulet, while ostensibly about creating a fact-checking tool in Google Sheets, offers something more valuable to developers: a clear framework for integrating powerful, multi-step Vertex AI Agents into Google Workspace solutions.
While the example is a FACT_CHECK custom function, the underlying architecture caught my eye as it provides a blueprint for connecting Google Workspace applications with the sophisticated reasoning capabilities of custom AI agents.
The Core Architectural Pattern
The solution uses Google Apps Script as a bridge between the user-facing application (Google Sheets) and the backend AI services on Google Cloud.
When a user calls the custom function, the script authenticates with a service account and makes a call to a deployed Vertex AI Agent. This agent then performs its multi-step reasoning. The result is then optionally passed to a Gemini model for final formatting before being returned to the Sheet.
This pattern is highly adaptable for various use cases, allowing developers to bring advanced Agentic AI into your own solutions.
Key Components for Developers
For developers looking to adapt this framework, the tutorial outlines the essential components:
A configured Google Cloud Project with the Vertex AI API enabled.
A deployed ADK Agent on the Vertex AI Agent Engine.
An Apps Script project using a service account for secure, server-to-server authentication with Google Cloud.
The provided Apps Script code (Code.gs and AiVertex.js) serves as a robust starting point, handling the API calls to both the reasoning agent and a Gemini model for final output formatting.
Ultimately, the fact-checking tool serves as an excellent proof of concept. The true value for the developer community, however, lies in the architectural blueprint it provides. This tutorial offers a clear model for integrating multi-step AI reasoning into Google Workspace add-ons, opening the door to a new class of intelligent applications.
This screenshot shows the new OAuth consent screen, which lets the user provide consent for a subset of the requested OAuth scopes. (Image credit: Google)
Earlier this year, we launched an improved version of the OAuth consent screen to the Apps Script IDE and unpublished Editor Add-ons that allows users to specify which individual scopes they would like to authorize for that script. For example, if a script requests access to a user’s Sheets and Forms files, and the users only intends to use the script with Sheets files, they can decide to only allow access to their spreadsheets and not their forms.
Google is continuing to enhance user privacy and control by extending the granular OAuth consent screen to published Editor Add-ons. This latest move builds upon the changes first introduced to the Apps Script IDE back in January, which we covered in detail at the time.
For those who missed it, this feature allows users to choose exactly which permissions (scopes) an application can access, rather than being forced into an “all-or-nothing” decision.
What’s Changing Now?
This more granular consent experience will now be presented to users when they install or re-authorize published Editor Add-ons from the Google Workspace Marketplace. When an add-on requests access, for instance to both a user’s Docs and Sheets, the user can now choose to approve one, both, or neither of those permissions.
While Google notes that “Pre-existing scope grants will not be affected”, it’s crucial for publishers of existing add-ons to understand that this new consent screen will appear if a user needs to re-consent. This can happen for several reasons, such as a user manually revoking the add-on’s access from their Google Account, or if a refresh token expires (e.g., after six months of inactivity). In these cases, existing users will be presented with the new granular options, and developers must be prepared for the possibility of partial authorisation.
A Reminder for Developers
As this functionality rolls out, starting on August 19 2025, it’s crucial for developers to ensure their add-ons can handle partial consent. If a user denies a specific scope that is essential for a feature, the add-on should fail gracefully without breaking.
As we detailed in our January post, the ScriptApp and AuthorizationInfo classes are your key tools for managing this. These allow you to programmatically check the granted scopes and build safeguards into your code. For a deeper dive into the methods and a guide on handling these scenarios, we recommend reviewing our original article.
In this video we build a Chat app as a Google Workspace add-on with Apps Script and extend it to other Workspace applications (Calendar, Gmail, Drive, Docs, Sheets, and Slides).
To see this new framework in action, Pierrick Voulet from the Google Developer Relations team has released a fantastic video tutorial, ‘Build a Chat app as a Google Workspace add-on with Apps Script’.
The ‘OOO Assistant’: A Practical Demonstration
The video centres on building an ‘Out-of-Office Assistant’ that can block out the user’s Calendar for the day and set a Gmail out-of-office auto-reply. The real magic is how this Chat-based logic extends into a full Workspace Add-on, available in the sidebars of Gmail, Calendar, Drive, Docs, Sheets and Slides. This is achieved with minimal code changes, primarily by configuring the appsscript.json manifest to include:
A homepageTrigger to display a UI card in the add-on sidebar.
universalActions that reuse the exact same functions (blockDayOut(), setAutoReply()) originally built for the Chat quick commands.
The video provides a complete overview of the process, and you can get hands-on with the shared code yourself. For more technical details, check out the official Extend Google Chat documentation.
The landscape of what’s possible within the browser is quietly undergoing a significant shift, and for Google Workspace Add-on developers, this could be a game-changer. Chrome’s AI mission is simple yet powerful: to ‘make Chrome and the web smarter for all developers and all users.’ We’re seeing this vision begin to take shape with the emergence of experimental, built-in AI APIs in Google Chrome, designed to bring powerful capabilities using models like Gemini Nano directly to the user’s device.
There is a growing suite of these on-device APIs. This includes the versatile Prompt API, specialised Writing Assistance APIs (like Summarizer, Writer, and Rewriter), Translation APIs (Language Detector and Translator), and even a newly introduced Proofreader API. For many existing Workspace Add-on developers, some of the more task-specific APIs could offer a relatively straightforward way to integrate AI-powered enhancements.
However, my focus for this exploration, and the core of the accompanying demo Add-on being introduced here, is the Prompt API. What makes this API particularly compelling for me is its direct line to Gemini Nano, a model that runs locally, right within the user’s Chrome browser. This on-device approach means that, unlike solutions requiring calls to external third-party GenAI services, interactions can happen entirely client-side. The Prompt API provides web applications, including Google Workspace Editor Add-ons, with an open-ended way to harness this local AI for text-based generative tasks.
To put the Prompt API’s text processing abilities through its paces in a practical Workspace context, I’ve developed a Google Workspace Add-on focused on text-to-diagram generation. This post delves into this demonstration and discusses what on-device AI, through the versatile Prompt API, could mean for the future of Workspace Add-on development, including its emerging multimodal potential.
Why This Matters: New Horizons for Google Workspace Developers
Using an on-device LLM like Gemini Nano offers several key benefits for Workspace Add-on developers:
Enhanced Data Privacy & Simplified Governance:Sensitive user data doesn’t need to leave the browser, meaning no external API calls are made to third-party servers for the AI processing itself, which is a huge plus for privacy and can simplify data governance including Google Workspace Marketplace verification and Add-on data privacy policies.
Potential for Cost-Free GenAI (with caveats!): Client-side processing can reduce or eliminate server-side AI costs for certain tasks. Remember, “Nano” is smaller than its cloud counterparts, so it’s best for well-scoped features. This smaller size means developers should think carefully about their implementation, particularly around prompt design to achieve the desired accuracy, as the model’s capacity for understanding extremely broad or complex instructions without guidance will differ from larger models.
Improved User Experience & Offline Access:Expect faster interactions due to minimise network latency.
The biggest takeaway here is the opportunity to explore new avenues for GenAI capabilities in your Add-ons, albeit with the understanding that this is experimental territory and on-device models have different characteristics and capacities compared to larger, cloud-based models.
Proof of Concept: AI-Powered Text-to-Diagram Add-on
To showcase the tangible possibilities of on-device text processing, the demonstrator Add-on (available in the Nano-Prompt-AI-Demo GitHub repository) focuses on a text-to-diagram use case:
Users can describe a diagram in natural language (e.g., “flowchart for a login process”).
The Add-on then uses the Gemini Nano API via the Prompt API to convert this text description into MermaidJS diagram code.
It also allows users to directly edit the generated MermaidJS code, see a live preview, and utilise an AI-powered “Fix Diagram” feature if the code has errors.
Finally, the generated diagram can be inserted as a PNG image into their Google Workspace file.
Nano Prompt API Demo
This example illustrates how the Prompt API can be used for practical tasks within a Google Workspace environment.
Under the Bonnet: Utilising the Chrome Gemini Nano Prompt API for Text
The Add-on interacts with Gemini Nano via client-side JavaScript using the LanguageModel object in the Sidebar.html file. I should also highlight that all of the Sidebar.html code was written by the Gemini 2.5 Pro model in gemini.google.com, with my guidance which included providing the appropriate developer documentation and this explainer for the Prompt API.
The Add-on’s core logic for text-to-diagram generation includes:
Session Creation and Prompt Design for Gemini Nano:A LanguageModel session is created using LanguageModel.create().
Generating Diagrams from Text: The user’s natural language description is sent to the AI via session.prompt(textDescription).
AI-Powered Code Fixing: If the generated or manually entered MermaidJS code has errors, the faulty code along with the error message is sent back to the model for attempted correction.
Given that Gemini Nano is, as its name suggests, a smaller LLM, careful prompt design is key to achieving optimal results. In this demonstrator Add-on, for instance, the initialPrompts (system prompt) play a crucial role. It not only instructs the AI to act as a MermaidJS expert and to output onlyraw MermaidJS markdown, but it also includes two explicit examples of MermaidJS code within those instructions.
Providing such “few-shot” examples within the system prompt was found to significantly improve the reliability and accuracy of the generated diagram code from text descriptions. This technique helps guide the smaller model effectively.
Navigating Experimental Waters: Important Considerations (and Reassurances)
It’s important to reiterate that the majority of AI APIs are still experimental. Functionality can change, and specific Chrome versions and flags are often required. I recommend referring to official Chrome AI Documentation and Joining the Early Preview Program for the latest details and updates.
Before you go updating your popular production Google Workspace Add-ons developers should be aware of the current system prerequisites. As of this writing, these include:
Operating System: Windows 10 or 11; macOS 13+ (Ventura and onwards); or Linux.
Storage: At least 22 GB of free space on the volume that contains your Chrome profile is necessary for the model download.
GPU: A dedicated GPU with strictly more than 4 GB of VRAM is often a requirement for performant on-device model execution.
Currently, APIs backed by Gemini Nano do not yet support Chrome for Android, iOS, or ChromeOS. For Workspace Add-on developers, the lack of ChromeOS support is a significant consideration.
However, Google announced at I/O 2025 in the ‘Practical built-in AI with Gemini Nano in Chrome’ session that the text-only Prompt API, powered by Gemini Nano, is generally available for Chrome Extensions starting in Chrome 138. While general web page use of the Prompt API remains experimental this move hopefully signals a clear trajectory from experiment to production-ready capabilities.
Bridging the Gap: The Hybrid SDK
To address device compatibility across the ecosystem, Google has announced a Hybrid SDK. This upcoming extension to the Firebase Web SDK aims to use built-in APIs locally when available and fall back to server-side Gemini otherwise, with a developer preview planned (see https://goo.gle/hybrid-sdk-developer-preview for more information). This initiative should provide a more consistent development experience and wider reach for AI-powered features.
A Glimpse into the Future: Empowering Workspace Innovation
On-device AI opens new opportunities for privacy-centric, responsive, and cost-effective Add-on features. While the demonstrator Add-on focuses on text generation, the Prompt API and the broader suite of on-device AI tools in Chrome offer much more for developers to explore
Focusing on Unique Value for Workspace Add-ons
It’s important for developers to consider how these on-device AI capabilities—be it advanced text processing or new multimodal interactions which support audio and image inputs from Chrome 138 Canary—can be used to extend and enhance user experience in novel ways, rather than replicating core Gemini for Google Workspace features. The power lies in creating unique, value-added functionalities that complement native Workspace features.
Explore, Experiment, and Provide Feedback!
This journey into on-device AI is a collaborative one and Google Workspace developers have an opportunity to help shape on-device AI.
Try It Out: Follow setup instructions to experience on-device AI with the demo, and consider exploring multimodal capabilities for your own projects by referring to the latest Early Preview Program updates.
Provide Feedback: Share your experiences either about the example add-on or through the Early Preview Program.
I hope you have as much fun working with these APIs as I have and look forward to hearing how you get on. Happy Scripting!
It has finally happened! Bot Bender 2.0 is now available for installation from the Google Workspace Marketplace. In this post, I will walk you through the steps needed to get to the marketplace.
We’ve shared a couple of journeys to Google Workspace Marketplace publication but this is the first I’ve seen for publishing a Google Chat App. For people who have gone through the process it’s probably a familiar story of back and forth with the Marketplace review team.
In this post Anton Shevchuk shares how his Chat App received multiple rejects due to issues with the location of the support/privacy pages. The moral of the story is to be persistent, and don’t hesitate to push back if you believe the reviewers are mistaken. The post from Anton also outlines the other steps involved, including connecting the Google Workspace Marketplace SDK and configuring app visibility settings.
In 2022, we made it easy to seamlessly access popular Education Technology tools directly in Google Classroom. We partnered with 20+ EdTech companies, including Kahoot!, Pear Deck, IXL, ReadWorks, and Nearpod, to build Google Classroom add-ons. These new integrations let educators and students easily find, use, and grade great content in their favorite EdTech tools without having to navigate to external websites and apps.
Today, we’re excited to make Classroom add-ons generally available to all developers.
Given the number of Apps Script users in EDU today’s announcement the Google Classroom add-ons now generally available for anyone to publish is probably welcome news. Whilst Google Classroom add-ons have been around for a couple of years they have historically only been created by a select few of Google partners.
Before jumping in something key to keep in mind is that while Apps Script offers a low-entry approach for most Workspace add-ons, Classroom add-ons require more technical expertise. They involve building a web application and integrating it with Classroom’s functionalities using the Google Workspace Marketplace SDK. Proficiency in web development languages like Python, Java, or Node.js is recommended and Google have published some example implementations for Python, Java and Node.js.
Whilst there are differences there are overlaps. For example like Workspace and Editor Add-ons you will need to create a Google Cloud project to publish to the Google Workspace Marketplace. As part of this you’ll have to also provide key information such as terms and conditions as well as going through the OAuth verification process.
The Google Developer documentation and other resources are linked from the source announcement.
Use Google Apps Script to create a custom menu that will work inside Google Sheets, Google Docs, Slides and Google Forms.
Tired of writing separate code for custom menus in Docs, Sheets, Slides, and Forms? This nifty Google Apps Script snippet by Amit Agarwal helps you build a universal menu that works across all these apps!
The solution is a small snippet which detects the context to see which of DocumentApp, SpreadsheetApp, FormApp, SlidesApp are available which then lets you switch to get the appropriate .getUI() method.
Check out the source post for the full code and streamline your add-on development!
The Slides Advisor project is an open source, Google Workspace Add-On that uses artificial intelligence (AI) technology to review and give feedback on your presentations, whenever you need it. The project uses Google’s Gemini API image and text processing features to analyze your Google Slides against a set of guidelines and lets you know how you are doing against those guidelines. Even better, you can customize those guidelines to follow your organization’s recommendations, or remind you to improve your presentations based on rules you define for yourself.
We’ve all been there: scrambling to cram too much information onto a single slide. But what if you had an AI assistant to give your presentations a once-over?
The Slides Advisor is a free, open-source Google Workspace Add-on for Google Slides that uses the power of AI to analyze your presentations. It checks your slides against customisable guidelines, helping you ensure they’re clear, concise, and visually appealing.
Everything you need to get started is covered in the source post by Joe Fernandez and Steve Bazyl. One of the highlights for me is seeing how you can setup a service account to access the Gemini AI API with Google Apps Script. The post also covers how you can modify the prompt you use to get different responses from Gemini.
If you want to publish your Google Apps Script project on the Google Workspace Marketplace, or if you already have an add-on or app on this marketplace and are using authorization scopes which are now restricted, you will have to go through a TIER 2 CASA security assessment.
If you would like an overview of the process from a developer’s perspective you are in luck as Kelig Lefeuvre (Product Engineer at Scriptit & Folgo) has shared a guide specifically with Apps Script developers in mind. The article includes a number of useful tips and information about the review process which you won’t find in the official documentation.
As part of the CASA process Kelig recommends using the option to you can submit bypass the Fortify scan and submit your own results. those results Kelig has also provided a second article with a step-by-step guide on ‘how to scan your Google Apps Script project for CASA’.
