Cost to build a Google Workspace add-on

Building a Google Workspace add-on in 2026 typically costs $5,000 to $150,000+ depending on scope. A simple Apps Script utility for your own domain runs $5,000 to $15,000. A custom-UI add-on with Cards API and a few Google APIs lands at $20,000 to $50,000. A full Marketplace SaaS with billing, OAuth verification, and security review runs $50,000 to $150,000+. The biggest hidden cost is calendar time: OAuth verification plus Marketplace review can add 4 to 12 weeks before a single paying user touches your add-on.

Why "Google Workspace add-on" means three different things

The phrase covers three very different products with three very different budgets. Founders often quote one and end up paying for another.

The first is an internal Apps Script utility. A script that runs inside your own Workspace tenant, automates a repeatable task, and never leaves your domain. No OAuth review, no Marketplace listing, no public users. This is the cheap end.

The second is a custom-UI Google Workspace add-on, built on the Workspace Add-ons SDK with Cards API. It shows up inside Gmail, Docs, Sheets, or Calendar as a side panel. It calls one or two Google APIs (Drive, Gmail, Docs) plus your own backend. Distributed publicly, so OAuth verification and Marketplace review apply.

The third is a full Marketplace SaaS: custom UI, multiple Google APIs (often including restricted scopes like gmail.readonly or drive), Workspace Marketplace billing integration, admin install for whole Workspace domains, and a security assessment if you touch restricted scopes. This is the expensive end, and almost all the cost lives in approval and compliance, not code.

What goes into a Workspace add-on

The actual surface area of a Workspace add-on, in rough order of effort:

• Apps Script vs Workspace Add-ons SDK choice. Apps Script is the fastest path: free, hosted by Google, one IDE in the browser. The Add-ons SDK lets you point at your own backend (Cloud Functions, a Node service, Vercel, anything) and gives you more control over UI, auth, and observability. Pick Apps Script for utilities under 500 lines. Pick the SDK for anything you'll charge money for.
• OAuth scopes and the verification gauntlet. This is where most of the budget goes, and almost none of the budget is dollars; it's calendar time. Pick the narrowest scope that works. gmail.send is sensitive (3 to 5 business days nominal, 2 to 6 weeks real). gmail.readonly and drive are restricted, which triggers a CASA security assessment, a paid third-party audit costing several thousand dollars and adding weeks to months.
• Marketplace listing review. Separate from OAuth review. Typically 2 to 6 weeks for a clean submission. Can stretch longer if Google has questions about your privacy policy, your demo video, or your branding.
• Google APIs you'll touch. Drive API, Gmail API, Docs / Sheets / Slides APIs, Calendar API, Admin SDK. Each has its own scope decision and its own quirks. The Cards API is the UI layer for Gmail-side panels and Docs side panels.
• Auth flow choice: JIT vs preauthorized. Just-in-time auth pops the consent screen the first time the user clicks the feature. Preauthorized auth bundles all scopes at install. Preauthorized is friendlier for admin domain installs; JIT is friendlier for individual users testing. You usually need both, configured per surface.
• Marketplace billing. Optional but expensive to add late. If you want to charge users through Google's billing rails (so admin can buy seats during install), you wire up the Workspace Marketplace billing API and conform to Google's pricing display rules. Adds 2 to 4 weeks of work and one more thing for review.

The first three line items are 60% of total spend on most Workspace add-on projects. The actual feature code is usually less than half the budget.

Cost breakdown by approach

This is where most cost-to-build posts hand-wave. The real choice is not "which firm do I hire," it's "which of these three add-on shapes do I actually need." Here's the honest breakdown.

A few things worth saying out loud about the table.

The full-time hire row is brutal for a single add-on. You're not hiring someone for a Workspace add-on; you're hiring them for a year of work, and the add-on is the first 12 weeks of it. If that math doesn't work, don't do it.

The agency row is the default option and it's usually the wrong default. Agencies price the build, not the verification waiting. They hand the OAuth review back to you halfway through and disappear when Google asks for a second demo video. Read the SOW for who owns the back-and-forth with Google's reviewers.

The freelancer row works great for the Apps Script utility tier and falls apart at the Marketplace SaaS tier. Verification, billing, and admin install are specialty knowledge.

Every engineer on Cadence is AI-native by default, vetted on Cursor, Claude Code, and Copilot fluency before they unlock bookings. For Workspace add-ons specifically, that matters because Apps Script and the Add-ons SDK have well-documented surfaces that AI tooling handles cleanly. A mid-tier engineer with Claude Code can scaffold a full Cards API surface in a day instead of a week. The same way the cost to integrate Claude API into your app drops sharply when the engineer uses Claude Code to write the integration, Workspace add-on build time drops 30 to 50% when AI-native is the baseline.

Feature-by-feature cost breakdown

Concrete numbers, in engineer-weeks at the mid tier ($1,000/week):

• Basic Cards API side panel in Gmail or Docs: 0.5 to 1 week. $500 to $1,000.
• OAuth flow with one sensitive scope (e.g. gmail.send): 0.5 week to wire, then 2 to 6 weeks of calendar time waiting on Google. $500 in build cost; the cost is calendar.
• OAuth flow with one restricted scope (e.g. gmail.readonly, drive): 1 week to wire, then 4 to 12 weeks of calendar time, plus a CASA security assessment that runs $3,000 to $15,000 from a third-party firm.
• Drive API integration (read + write files): 1 to 2 weeks. $1,000 to $2,000.
• Docs / Sheets / Slides API edits: 1 to 3 weeks per surface. The APIs are powerful but verbose. $1,000 to $3,000 per surface.
• Marketplace listing: 0.5 week to assemble (privacy policy, terms, demo video, OAuth justification, branding). $500. Plus 2 to 6 weeks calendar wait.
• Marketplace billing integration: 2 to 4 weeks. $2,000 to $4,000.
• Admin install + domain-wide delegation: 1 to 2 weeks of testing and edge-case handling. $1,000 to $2,000.
• CASA security assessment (only if you have restricted scopes): 4 to 12 weeks, $3,000 to $15,000 in audit fees, plus engineer time addressing findings. Budget another $3,000 to $8,000 in remediation work.

A useful shortcut: tally your scopes first. If you're using only basic and sensitive scopes, you're in the $20,000 to $50,000 build tier and you'll be in market in 6 to 10 weeks. Add a single restricted scope and you've moved into the $50,000 to $150,000+ tier and a 4-month calendar window.

How to reduce costs without cutting corners

Five practical levers, ranked by savings:

1. Avoid restricted scopes if you possibly can. This is the single biggest lever. If you can do your add-on with gmail.send, gmail.compose, or gmail.addons.execute instead of gmail.readonly, you skip the CASA assessment, save $5,000 to $15,000, and shave 4 to 8 weeks off your launch. Read the Workspace add-ons scopes guide carefully and design the feature around the narrowest scope that works.
2. Start with Apps Script even if you'll outgrow it. A 3-day Apps Script prototype proves the workflow, gets internal users typing, and gives you a real spec for the SDK rebuild. Most teams skip this and spend an extra 2 weeks designing in the abstract.
3. Use the Workspace Add-ons SDK only when Apps Script truly can't do the job. If your add-on calls only Google APIs and lives entirely in the side panel, Apps Script is fine and free to host. Move to the SDK when you need your own backend, custom auth, or third-party API calls that exceed Apps Script's quotas.
4. Bundle your verification submission. Submit OAuth verification, Marketplace listing, and CASA (if applicable) in parallel where Google's process allows. Submitting them serially adds weeks of dead calendar.
5. Use AI-native engineers for the boilerplate. Cards API code, Apps Script auth wrappers, Drive API CRUD code: this is exactly the territory where Cursor and Claude Code outperform manual coding. The same logic that makes the cost to build a Chrome extension drop sharply with AI-native tooling applies here. Roughly the same shape of API, same scope-vs-permission gauntlet.

The fastest path from idea to Workspace add-on

Three steps, in this order:

1. Spend a day asking whether you actually need to build this. Search the Google Workspace Marketplace for your use case. Then check whether a Zapier or Make integration covers it. Roughly 70% of "we need a Workspace add-on" requests we see at Cadence are met by an existing Marketplace listing plus a no-code automation. If yours is one of them, your cost just dropped from $20,000 to $0.
2. If you do need custom: scope your OAuth before you scope your features. List every Google API call your add-on will make, map each to its scope, and confirm whether you're in the sensitive-only path or the restricted path. The scope decision determines your timeline and your audit budget. If you're in the restricted path, decide right now whether the feature is worth 4 months of calendar time and a $5,000 audit. Often it isn't.
3. Book a mid or senior Cadence engineer with Workspace SDK experience and run the 48-hour trial on the riskiest part of the project. That's usually the OAuth flow plus a working Cards API side panel calling one Google API end to end. If you can ship that in two days, the rest is execution. If not, you have data to renegotiate scope before spending real money. You can see what it costs on Cadence or run the numbers on the engineering ROI calculator first.

Cadence books a vetted, AI-native engineer in 2 minutes from a 12,800-engineer pool, with a 48-hour free trial. Weekly billing means you're not married to a Workspace add-on project that turns out to be a Marketplace listing in disguise. Replace any week, no notice.

The dirty secret of Workspace add-on economics: most of the cost is verification waiting, not coding. Pick a scope you can verify in two weeks, not two months. Pick an engineer who's shipped a Workspace add-on through Google review before, not someone learning on your project. And ask the build-vs-buy question before you sign the SOW. Most "we need a Workspace add-on" projects don't, and the ones that do are mostly an OAuth submission with some glue code attached.

For the broader question of which AI tools your engineer should bring to this work, the best AI coding tools 2026 review walks through the current Cursor / Claude Code / Copilot tradeoff. The answer for a Workspace add-on is "all three, used for different parts."

FAQ

How long does it take to build a Google Workspace add-on?

A simple Apps Script utility takes 1 to 3 weeks. A custom-UI Workspace add-on with sensitive OAuth scopes takes 6 to 10 weeks total (build plus verification plus listing review). A full Marketplace SaaS with restricted scopes and CASA security assessment takes 12 to 20 weeks. The build itself is usually under half the calendar time; OAuth verification and Marketplace review dominate.

Do I need a Google Workspace subscription to build an add-on?

No, you can develop and publish a Workspace add-on with a free Gmail account and Google Apps Script. You'll want a paid Workspace account for testing admin install flows and domain-wide delegation. That's $7 to $22 per user per month, not a meaningful line item against the build cost.

What's the difference between Apps Script and the Workspace Add-ons SDK?

Apps Script is Google's hosted JavaScript runtime: free, in-browser IDE, runs on Google's infrastructure, has quotas. The Workspace Add-ons SDK lets you build the same add-on UI but point it at your own backend (Cloud Functions, a Node server, Vercel, anything). Apps Script is the right choice for utilities and prototypes; the SDK is the right choice for anything you'll charge for or scale beyond Google's quotas.

Does Google charge a fee to publish on the Workspace Marketplace?

No. Listing on the Google Workspace Marketplace is free. The cost is engineer time to assemble the listing (privacy policy, terms, demo video, OAuth justification, branding) plus calendar time waiting for review (2 to 6 weeks typical).

When do I actually need a CASA security assessment?

Only when your add-on requests restricted OAuth scopes, mostly gmail.readonly, gmail.modify, drive, and drive.readonly. CASA is a paid third-party audit ($3,000 to $15,000) and adds 4 to 12 weeks. Designing your add-on to use sensitive scopes only (gmail.send, gmail.compose, drive.file) avoids it entirely. This single design decision is the biggest cost lever in the whole project.

Build vs Buy vs Hire: how do I decide for a Workspace add-on?

Search the Marketplace first; if a listing covers 80% of your need, buy it and bridge the rest with Zapier. Build custom only when the workflow is specific to your customer base (e.g. a vertical CRM that lives inside Gmail for legal practices) and no Marketplace listing comes close. Hire full-time only if you'll have 6+ months of related work after launch; otherwise book on-demand for the build window and reassess after launch.