How to use Google Opal to build client-ready AI tools in under an hour

Imagine handing a client a fully working AI tool built just for their business—without writing a single line of code, paying a developer, or buying a software subscription. That’s exactly what Google’s new Opal builder makes possible.

Opal is quietly one of the most powerful free no-code AI app builders available right now, and it’s already integrated directly into Gemini. If you sell services, consult, freelance, or run a small agency, this is the kind of leverage that can change your entire offer.

What is Google Opal?

Opal is a free no-code app builder released through Google Labs. Instead of writing code, you describe the tool you want in plain language. Opal then turns that description into a working mini-application powered by Gemini.

Behind the scenes, Opal automatically:

• Maps out the workflow your app needs
• Chooses the right Gemini models for each step
• Builds the user interface
• Connects inputs, processing, and outputs into a live, shareable app

The result is a real, usable tool you can hand to a client—without touching code or managing APIs.

How Opal connects to Gemini

Opal isn’t a separate product you have to hunt down. It now lives inside the Gemini web app as part of the "gems" experience.

The flow is simple:

1. Open Gemini in your browser.
2. Go to the gems section.
3. Create a new gem and describe what you want your mini app to do.
4. Gemini’s latest models (including Gemini 3.1 Pro) build and run the logic for your app—completely free.

This tight integration means your apps are powered by the same cutting-edge models you’d normally access through paid platforms or APIs, but wrapped in a no-code builder.

The “deployed asset” model: why this matters for service businesses

Most freelancers and small agencies are stuck selling time: reports, campaigns, documents, and other manual deliverables. Opal lets you shift to what you can think of as the “deployed asset” model.

Instead of doing the same work over and over, you:

• Build a reusable AI tool once
• Deploy it into the client’s business
• Let it keep producing value without you being there

This turns your work into an asset that runs in the background—something clients can justify paying for setup and ongoing maintenance, not just a one-off deliverable.

If you like this way of thinking, you’ll probably also get value from a broader roadmap like how to build an AI-powered app in 2026, which expands on turning AI tools into real products.

Two ways to build apps in Opal

When you open the Opal builder, you have two main ways to create:

1. Conversational building

This is the easiest starting point. You simply describe what you want the app to do in natural language. Opal then auto-generates the full workflow:

• Input steps (what the user provides)
• Processing steps (what Gemini does with that input)
• Output steps (what the user gets back)

For most people, this is the fastest way to go from idea to working tool.

2. Visual node editor

If you want more control, you can switch to a visual editor where each part of your app is a node you can drag, drop, and connect.

Each node represents a specific action, such as:

• Input nodes – collect details from the user
• Processing nodes – run Gemini prompts, analyze, transform, or generate content
• Output nodes – display results, summaries, or media

You can chain these nodes in any order to create more complex workflows, and preview everything in real time as you build.

Key features that make Opal feel like real software

Opal isn’t just a fancy prompt wrapper. Several features push it into “real app” territory—good enough to charge for.

Session-level memory

Opal supports persistent memory at the session level. That means your app can remember important details across uses, such as:

• A client’s preferred tone or communication style
• Their target market or niche
• Key constraints or preferences

When users return, the app doesn’t start from zero—it builds on what it already knows. This feels far more professional and is exactly the kind of polish clients will pay recurring fees to keep.

Dynamic routing and branching logic

Real users don’t always follow a perfect path. Opal handles this with dynamic routing:

• Your workflow can branch based on user answers
• The app can choose different paths depending on conditions
• It can adapt in real time instead of running a rigid script

This is how you build intelligent agents instead of brittle automations that break when something unexpected happens.

Interactive pauses and clarifying questions

Another powerful feature is the interactive pause. Your app can:

• Stop mid-workflow
• Ask the user a clarifying question
• Wait for their response
• Then continue with better information

This dramatically improves output quality and reduces the “generic AI response” problem, especially for nuanced tasks like briefs, summaries, or strategy suggestions.

Built-in media and web capabilities

Opal apps can do much more than text generation. Inside a single workflow, your tool can:

• Pull live information from the web
• Generate images
• Produce short videos
• Combine research, writing, and visuals into one output

All of this works without extra APIs or third-party tools. That’s the kind of capability that used to require multiple paid platforms stitched together.

Clean delivery: shareable links and easy embedding

Every Opal app you publish generates a shareable link. That link is your delivery mechanism.

You can:

• Embed the app on a client’s website
• Add it to a client portal
• Drop it into an email or internal wiki

Wherever the link lives, the app runs. Clients don’t have to learn a new platform or log into a complex dashboard—they just click and use the tool. That simplicity makes it much easier to sell and onboard.

Real-world examples you can build and sell

To see the business potential, think about service businesses that are currently paying for tools or doing repetitive work manually. Each of these can be turned into an Opal app:

• Personal injury law firm – A case intake tool that asks structured questions, summarizes the case, and outputs an attorney-ready brief.
• Fitness studio – A custom workout plan generator that coaches use to create tailored routines based on goals, experience, and limitations.
• Property management company – A maintenance request assistant that categorizes issues, prioritizes them, and drafts response emails automatically.
• Marketing consultant – A competitive analysis tool that researches a brand’s rivals and outputs a formatted report with key insights.

In each case, you can charge for:

• Initial build and setup
• Monthly maintenance, updates, and improvements

Opal does the heavy technical lifting. Your value is in understanding the business problem, designing the workflow, and delivering a tool that fits neatly into the client’s existing process.

A simple starter scenario for freelancers

Here’s a concrete example of how a solo freelancer might use Opal.

Imagine a content freelancer who works with health and wellness coaches. Her clients constantly ask for:

• New client intake forms
• Session summaries
• Follow-up emails

She spends a single Saturday building an Opal app that:

1. Collects a new client’s health goals, habits, and restrictions.
2. Runs that information through a Gemini-powered processing step.
3. Outputs a clean client profile plus a suggested first session outline.

After a couple of test runs and prompt tweaks, she now has a reusable tool she can bundle into every new client engagement. It feels like custom software, cuts her delivery time, and strengthens her positioning—without any coding.

If you want to go deeper on this kind of "build once, sell many" approach, you may also like this guide on building a profitable AI app in minutes with no code, which follows a similar pattern in a different niche.

Using Opal’s template gallery

If you’re not sure where to start, Opal includes a template gallery maintained by Google. These templates cover different use cases and come with pre-built workflows you can study and adapt.

You can:

• Open a template that’s close to your client’s use case
• Inspect how the nodes are connected
• Adjust prompts, inputs, and outputs to match your niche
• Have a working foundation in minutes instead of starting from a blank canvas

This is especially helpful if you’re new to thinking in workflows or don’t yet have a strong mental model for how to structure AI-powered tools.

What your business could look like 30 days from now

Opal changes the starting position for anyone offering services. In 30 days, you could realistically have:

• Three different Opal apps deployed for three different clients
• A new line item in your offers for "custom AI tools" or "internal AI assistants"
• Recurring revenue from maintenance and upgrades

Instead of only selling your time, you’re delivering assets that run inside your clients’ businesses every day. And because Opal is free and no-code, the main investment is your creativity and understanding of client problems—not software spend or development hours.

If you’ve been waiting for a low-risk way to step into AI productized services, Google Opal is exactly that. Open Gemini, start a new gem, describe the tool you wish existed for your best client, and build something they can actually use this week.