Mozilla's rules, read closely
Firefox Extension Development: Porting a Chrome Add-on, Getting It Signed by Mozilla, and the Manifest V2 Question
Firefox is the browser where Manifest V2 still runs. That makes it the most forgiving home for an old extension and, oddly, the one that trips up Chrome developers most, because Firefox's version of Manifest V3 is not quite Chrome's.
November 17, 2026 · 8 min read
- manifest_version 3 read, extension ID notes@acmefieldnotes.com
- background.service_worker is not supported
- background.scripts added beside service_worker
- Load Temporary Add-on: extension running
- data_collection_permissions declared for AMO
- readable source and build steps attached to submission
The short answer
A Firefox extension uses the same WebExtensions APIs as Chrome, packaged with Firefox's own manifest entries: an extension ID, background scripts instead of a service worker, and, for new AMO submissions since November 3, 2025, a data collection declaration. Mozilla signs every extension through addons.mozilla.org, listed or self-distributed. DataCram builds and ports them from $1,500, depending on complexity.
On this page
- The one major browser still running Manifest V2
- Five manifest and policy changes between a Chrome build and a Firefox build
- Signing: Mozilla's stamp, whether the add-on is listed or not
- Our route from a working Chrome extension to a signed Firefox one
- GatorCrawl Keywords, if it moved to Firefox
- Skip Firefox for now if any of these is true
- What a Firefox build costs, and whose account it lives in
The one major browser still running Manifest V2
Chrome finished with Manifest V2. As of August 31, 2026, Google's support timeline records all remaining Manifest V2 extensions removed from the Chrome Web Store. Firefox took the other road. Mozilla wrote in a March 2024 post that "Firefox, however, has no plans to deprecate MV2 and will continue to support MV2 extensions for the foreseeable future," and that it keeps supporting "the blocking webRequest feature." We have not read a newer statement from Mozilla, so that is where the record stands.
Firefox supports the new format too. Mozilla's migration guide says "Manifest V3 became generally available in Firefox 109 after being available as a developer preview from Firefox 101." The catch is in the details. Firefox runs background code as event pages ("non-persistent background pages", in Mozilla's words) for both versions, not as the service worker Chrome requires. A developer who hears "Manifest V3" and assumes it means the same file in both browsers loses a morning finding out otherwise.
One rule is identical, and older than Chrome's. The same guide notes that "Mozilla’s long-standing add-on policies prohibit remote code execution." Everything the extension runs ships inside the package. Our Manifest V3 guide explains why Chrome adopted the same rule.
Five manifest and policy changes between a Chrome build and a Firefox build
Porting is mostly additive. You keep the Chrome code and add what Firefox asks for. MDN's Chrome incompatibilities page removes one old chore: "Support for the browser namespace and promises are no longer a source of incompatibility." Here is what remains, from Mozilla's documentation as read in November 2026.
| Change | What Mozilla says | What we do |
|---|---|---|
| Background code | "background.service_worker is not supported"; for cross-browser Manifest V3, "specify both scripts and service_worker in the background key" | List both, and write the background code so it behaves under either |
| Extension ID | Mandatory for signing Manifest V3 extensions: "AMO does not assign an ID" | Set an ID in browser_specific_settings, in the form name@yourdomain, and never change it |
| Data collection | browser_specific_settings.gecko.data_collection_permissions must be provided for new AMO submissions from November 3, 2025 | Declare what the extension sends, consistent with what the privacy policy says |
| Source code | Minified or machine-generated code is allowed, but "Mozilla needs to review a copy of the source code" first; obfuscation is not allowed | Submit the readable source and build steps alongside the package |
| Remote code and transport | "Add-ons must be self-contained and not load remote code for execution" and "must use encryption when transporting data remotely" | All logic in the package; every call to your server over HTTPS |
Sources: MDN on the background key and browser_specific_settings, and Mozilla's add-on policies. None of the five is hard. All five are easy to forget, and each one is a rejection waiting for a Friday.
Signing: Mozilla's stamp, whether the add-on is listed or not
Every Firefox extension goes through Mozilla, even one you never list. The signing overview is plain about it: "Extensions and themes need to be signed by Mozilla before they can be installed in release and beta versions of Firefox," and signing "takes place through addons.mozilla.org (AMO), whether you choose to distribute your add-on through AMO or to do it yourself." Self-distributed add-ons are "sometimes referred to as “unlisted” extensions because they cannot be publicly viewed or installed from AMO."

On timing, Mozilla's own words, read in November 2026: "all add-ons undergo automated validation before they are signed. It can take up to 24 hours for your submission to be signed and published, or longer if your submission is selected for manual review." And: "All add-ons are subject to a manual code review at any time after submission." That last sentence is the reason we submit readable source and a tidy permission list, the same discipline our permissions and privacy page describes for Chrome. A review can come long after launch, and it should find nothing to ask about.
Our route from a working Chrome extension to a signed Firefox one
- 1
Read the Chrome build
List every API, permission and site the extension uses, and flag anything Firefox handles differently before quoting.
- 2
Extend the manifest
Add background scripts beside the service worker, the extension ID and the data collection declaration.
- 3
Load it as a temporary add-on
In about:debugging, "This Firefox", then "Load Temporary Add-on". Mozilla notes it "stays installed until you remove it or restart Firefox", which keeps test runs honest.
- 4
Test the paths that matter
Sign in, the main job on each target site, settings saved and restored, and the background waking up after Firefox has been idle.
- 5
Package source with the build
If anything is bundled or minified, the readable source and build instructions go with the submission.
- 6
Submit under your AMO account
Listed for the public, or self-distributed for staff and partners, signed by Mozilla either way.
Most of the time goes into the fourth step. Firefox's event pages wake and sleep on their own schedule, and an extension that keeps important state in a variable instead of storage can behave in one browser and misbehave in the other. We write state to storage from the first line, which also suits Chrome's service worker rules. The cross-browser extensions page explains how one codebase serves both.
GatorCrawl Keywords, if it moved to Firefox
DataCram's own extension makes a useful test. GatorCrawl Keywords is a keyword research web app, and its Chrome extension shows search volume, CPC, difficulty and a 12-month trend on Google's results pages. The manifest declares its background only as a service worker, src/background.js. That one line is the first thing a Firefox port changes.
- Background: add a scripts entry pointing at the same file, so Firefox runs it as an event page and Chrome keeps its service worker.
- State: the service worker already keeps lookups for an hour in session storage rather than in variables, so a sleeping background loses nothing.
- Extension ID: add one in browser_specific_settings, since AMO will not assign it.
- Data collection: the Chrome listing declares personally identifiable information, authentication information and website content. The Firefox declaration has to tell the same story.
- Testing: storage, the right-click menu, messaging and the web app's connect page each get checked in Firefox before an estimate is written.
The rest is already in Firefox's favor. The service worker's header comment reads "No analytics, no tracking, no browsing history." Its content script runs on one address pattern, and its host permission names one site, the GatorCrawl app. A reviewer has little to puzzle over. Read the full build in how we built the GatorCrawl extension.
Skip Firefox for now if any of these is true
We would rather lose a small job than sell one you do not need. A Firefox build earns its cost when your users choose Firefox. It wastes money when they do not.
- Your extension is for your own staff, and IT manages Chrome or Edge for everyone. The internal team extensions page covers that route.
- You have not yet seen whether the Chrome version earns installs. Prove the idea on one browser first.
- Your Chrome extension is still on Manifest V2 and dying there. Moving it to Firefox to avoid the work only delays it, since Mozilla's statement dates from March 2024 and could change.
- The job could be a web page or a feature inside your app. The extension versus web app comparison is the cheaper conversation.
What a Firefox build costs, and whose account it lives in
Custom browser extensions from DataCram start at $1,500, depending on complexity, with a fixed written estimate before work starts. A Firefox port of an existing Chrome extension is quoted the same way, after we read the code. Store submission, privacy disclosures and the permission plan are part of the build. See the extension pricing page for what moves the number, and custom Chrome extension development for how a build runs from discovery to handover.
You own the code, and we recommend the extension be signed and listed under your own AMO account, so updates never depend on us. Connections to your own systems, a CRM, a web app or an AI model, go live the same day through API integrations. Updates after launch are quoted; the extension maintenance page covers what Firefox and Chrome releases ask of you. DataCram is a web studio in San Antonio, Texas, building for the web since 1999. Browse our work or call (210) 346-0848. The full section starts at custom browser extensions.
Questions owners ask about custom browser extensions
Mozilla said in March 2024 that Firefox had no plans to deprecate Manifest V2 and would support MV2 extensions for the foreseeable future, including blocking webRequest. We have not seen a newer Mozilla statement. Firefox also supports Manifest V3, generally available since Firefox 109.
Usually not without changes. MDN says Firefox does not support background.service_worker, so the manifest needs background scripts too, and Manifest V3 extensions need an extension ID in browser_specific_settings because AMO does not assign one. New AMO submissions since November 3, 2025 also need a data collection declaration. AMO does accept .crx packages.
Yes. Mozilla's signing overview says extensions must be signed by Mozilla before release and beta versions of Firefox will install them, and signing happens through AMO whether you distribute there or yourself. Self-distributed add-ons are often called unlisted, because they cannot be viewed or installed from AMO.
Mozilla says it can take up to 24 hours for a submission to be signed and published, or longer if it is selected for manual review, and that any add-on can get a manual code review at any time after submission. We do not promise a time. We do make the submission easy to review.
Mozilla's add-on policies allow minified or machine-generated code but say Mozilla needs to review a copy of the source before those steps were applied. Obfuscated code is not allowed. If your extension is bundled, the readable source and build instructions go in with the submission.
The Mozilla pages we read on signing, submission and add-on policies state no developer fee, as of November 2026. Check AMO when you register. Our build price starts at $1,500, depending on complexity, and covers submission under your account.
Keep reading
- One build, four browsersCross-Browser Extension Development: One Codebase for Chrome, Edge, Firefox and Safari, and the Four Places It SplitsOne browser extension for Chrome, Edge, Firefox and Safari: what carries over, what splits, store fees as of November 2026. Built from $1,500.
- The format Chrome keptManifest V3 Migration: What Changed, the Manifest V2 Shutdown Date by Date, and What It Means for an Older ExtensionManifest V3 explained: service workers, no remote code, declarativeNetRequest, and Chrome's Manifest V2 phase-out dates exactly as Google lists them.
- An extension inside an appSafari Web Extension Development: Converting a Chrome Extension With Apple's Packager, Xcode and the App StoreSafari web extensions built new or converted from Chrome: Apple's packager, the Xcode app wrapper, App Store review and the program fee. From $1,500.
- The offer, part by partCustom Chrome Extension Development Services: Discovery, a Permission Plan, the Build and a Listing in Your NameHire a Chrome extension developer for discovery, a written permission plan, a Manifest V3 build, store submission and a handover under your account.
- Your app, one tab closerSaaS Browser Extensions That Carry Your App's Login, Plan and Credits Into the Tab Your Users Already Have OpenA companion Chrome extension that signs in with your SaaS account, spends the same plan credits and bills on your own site. From $1,500, owned by you.
- Your data, on their pageCustom SEO Chrome Extensions That Paint Your Own Numbers Onto Google's Results PageBuild an SEO extension that shows your own keyword, rank or client data on Google results, on one site, with two permissions. From $1,500, yours to own.
- Everything in Custom Browser ExtensionsSee the section
From the Insights desk
- 7 Signs Your Website Is Quietly Losing You CustomersA website rarely fails loudly — it just quietly lets customers slip away, and you never see the ones who left. Seven signs yours is leaking, from slow mobile loads to stale listings to no conversion tracking, and the fix for each.
- AI Agents vs. Plugins: Why Athena CMS Hires AgentsA plugin adds a feature and waits for someone to use it. An agent has a name, one job and a schedule, and it does the work. Seven reasons Athena CMS hires agents instead of installing plugins, where a plugin is still the right call, and how to pick your first agent.
Get your extension ready for Mozilla's review
Send us the Chrome extension or the idea. Someone at DataCram replies with what Firefox will need changed, what we would test, and a fixed written estimate.
Which of these sounds like you? (pick any)
Rather talk now? Call (210) 346-0848.