SmartProxy's newest tag says 2.4.2 while the release it names says 2.4.3
Firefox/Chrome browser extension. SmartProxy will automatically enable/disable proxy for the sites you visit, based on customizable patterns.
At a glance
- What is it?
- A browser extension that decides per site whether your traffic goes through a proxy, shipped to Firefox, Chrome and Edge from one webpack configuration with eight build targets. The details worth reading are not in the feature list: a git tag that disagrees with its own release name, three TypeScript packages under two names, a geolocation database kept out of the tree for licensing reasons, and a comparison with older tools that carries no evidence behind it.
- Who is it for?
- This is a reasonable extension for someone who already has proxy servers and wants per site routing without touching settings by hand, and it is a poor fit for anyone expecting a proxy service, since the file describes no plans, pricing or bandwidth. Three things to check before trusting a build.
- Can I use it commercially?
- Yes, with conditions. GPL-3.0 is a copyleft licence: if you distribute software that includes it, you must release that software's source code under the same licence. Running it internally without distributing it does not trigger that obligation.
- Is it still maintained?
- Yes. The repository received new commits within the last day.
- What is it written in?
- Mainly TypeScript, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.
Star history
- salarcode/SmartProxy2,774 ★ · +19
| Repository | 2026-10-04 | 2026-11-15 |
|---|---|---|
| salarcode/SmartProxy | 2.8k | 2.8k |
19 Stars in the last 42 days, measured from the daily GitHub snapshot. Compare with other projects
Editorial analysis
Tag v2.4.2 ships as Version 2.4.3
The three releases carry three different kinds of naming. The newest git tag is v2.4.2, published on 16 September 2026, and the release it names is titled Version 2.4.3. The tag therefore sits one increment below the version it points at. The release before it is v2.3 tagged with the name Version 2.3, which agrees with itself. The third is v2.2 carrying the name Version 2.2 + Version 2.2.1, which is two releases folded into a single entry. Only the newest row is inconsistent, and it is the row a user installing by tag meets first: clone v2.4.2 and the code inside is Version 2.4.3. Nothing in the repository reconciles the two numbers, and the manifest at the root carries no version field at all, so the version string inside a built extension comes from a file this one does not show. The last push to the default branch is dated 18 September 2026, two days after that release.
Three TypeScript packages under two names
The devDependencies carry three compiler entries, and the names a developer would type are not the names npm installs:
"typescript": "npm:@typescript/typescript6@^6.0.2",
"@typescript/native": "npm:typescript@^7.0.2",
"tsc": "^2.0.4",The familiar name typescript is an npm alias pointing at a scoped package called @typescript/typescript6 on the 6.x line. The scoped name @typescript/native is an alias pointing back at the plain typescript package on the 7.x line. The third, tsc, is a separate package from before the compiler was bundled. Two of the three are therefore the same compiler under two names at two different major versions, and which one the webpack build resolves depends on what ts-loader picks up. Nothing in the manifest records that choice. The test side has a comparable gap: jest sits at 30.5.1 while ts-jest is at 29.4.12, so the transformer is a full major version behind the runner it plugs into. There is also no dependencies block at all, only devDependencies, which fits an extension that bundles its whole surface into the build directory.
Eight build targets and three store links
The scripts block defines one webpack invocation per target, and the matrix is wider than the download links. There is build for plain chrome, build-ch which adds service_worker=true, build-ch-mv2 whose target is named for the previous manifest generation, build-ff, build-ff-unlisted, build-ff-android, build-ed, and build-op, with a watch variant beside each that also sets dev=true. Chrome is the interesting case because it has two separate artifacts rather than one, and which installations accept which of them is decided by the manifests those flags produce. Opera has a build target and no download link anywhere in the file, while the store links cover Firefox on desktop and Android, the Chrome Web Store, and the Edge Add-ons site. Read the list bottom up and the last entry visible is the Opera watch target, whose flag list stops partway through, so the scripts the file shows end there.
The geolocation database is kept out of the tree
The attribution note is the shortest section in the file and it has the largest practical consequence. SmartProxy uses IP2Location LITE for IP geolocation, and the database is not included in the repository because of licensing requirements. That means the checkout cannot produce a complete shippable extension by itself: whatever the interface needs to know about where an address sits arrives from outside the tree, at build time or at run time, and the file does not say which. It also means the licensing is not uniform. A GPL-3.0 LICENSE sits at the root of the repository, and the attribution names a separate data product under its own terms, so anyone redistributing a build holds one licence for the code and a second condition for the data. The three store links near the top are the route most users will take instead, and going that way leaves the question of how the data got into the package entirely to the store build.
The Firefox instructions end with any file inside the extension
Two loading procedures are given, both written for someone who already has a working build. The Firefox route is five steps: build with npm run build-ff, open Firefox, enter about:debugging in the URL bar, click Load Temporary Add-on, and then open the extension's build directory and select any file inside the extension. That last clause is doing a lot of work, because the steps never name which file, only any file inside the extension. The Chrome route is also five steps: build with build-ch, navigate to chrome://extensions, enable Developer Mode with the toggle switch, click LOAD UNPACKED, and select the extension's build directory. The Chrome steps name a directory and stop there. The Firefox steps hand the chooser to the developer. The stated requirements for either route are Node.js and a Firefox or Chrome browser, and the first two commands in the file are npm install and npm run build-ff:watch.
Travis config and a PowerShell packer sit beside the webpack build
The root mixes several generations of tooling. The current part is webpack.config.js driven by npm scripts, with ts-loader for TypeScript and clean-webpack-plugin, copy-webpack-plugin and terser-webpack-plugin around it. Beside that sits .travis.yml, a configuration file for a continuous integration service that none of the npm scripts reference and that nothing in the file explains. There is also build-and-pack.ps1, a PowerShell script, which means part of the packaging path runs only on a Windows machine even though the extension targets four browser families. Two further root files sit alongside: updateinfo.json, in the shape Firefox reads for updates, and vwd.webinfo. Linting has its own configuration in .eslintrc.json with an .eslintignore beside it, formatting is in .prettierrc, translations are wired through crowdin.yml with a wiki/ directory for documentation, and TypeScript settings live in tsconfig.json.
The routing decision is a domain pattern plus one global switch
Read the highlights as a description of how a single request is decided. Patterns are matched against the site being visited, and a match either turns the proxy on for that site or leaves it off, so the normal behaviour is a per domain router rather than a global one. Two entries cut against that shape. There is a mode for enabling proxy for all domains, and there is a view over the current page's items and requests where each one is proxified with a single click, which puts the decision below the domain level on individual requests. Two more entries concern where the proxy list comes from rather than how it is applied: the proxy API is fed through subscriptions, and there is backup and restore for settings and rules. The file says both exist and describes neither. Where a subscription list is kept, what a backup file contains, and whether restoring it overwrites a live configuration are all left to the wiki/ directory and the store pages.
The comparison with AutoProxy arrives without a measurement
The file names its predecessors in one sentence: SmartProxy is inspired by AutoProxy and AutoProxy-ng, says those are now considered legacy, and then claims to be completely written from the ground up using WebExtensions with more functionality than those proxies. Three claims share that sentence and none is supported. Nothing lists what the older tools do that this one does not, nothing dates the legacy judgement, and no measurement of any kind appears anywhere in the repository. A reader is asked to accept the comparison on the strength of the assertion itself. The highlight list is the only inventory on offer, and it is a list of intentions rather than a list of differences, so it cannot settle the question it is being used to settle. For anyone weighing the three, the useful comparison would sit on the store pages or in the wiki/ directory, and the file links to neither next to the claim.
Editorial conclusion
This is a reasonable extension for someone who already has proxy servers and wants per site routing without touching settings by hand, and it is a poor fit for anyone expecting a proxy service, since the file describes no plans, pricing or bandwidth. Three things to check before trusting a build. Confirm which Chrome artifact you are installing, because the manifest produces both a service worker build and one named chrome-mv2. Expect the geolocation data to arrive from outside the checkout, since the IP2Location database is excluded for licensing, so a build from a bare clone is not the same artifact as the store build. And pin by the version inside the extension rather than by the tag, given that the newest tag is one increment below the release it names. Anyone extending the source should read the manifest first: the compiler entries are aliased, the test transformer trails the test runner by a major version, and the visible script list ends partway through the Opera watch target.
Frequently asked questions
What does the SmartProxy extension do?
It matches patterns against the site you are visiting and switches the proxy on or off for that site, so you do not change proxy settings by hand. It also offers a mode that proxies all domains, a per-request view, proxy API fed by subscriptions, and backup and restore for settings and rules.
Is SmartProxy a residential proxy provider?
No. This repository holds a GPL-3.0 browser extension for Firefox, Chrome, Edge and Opera whose job is switching the browser's proxy setting per site through pattern matching. Nothing in the file describes selling bandwidth, subscription plans, pricing or affiliate programs.
Which browsers does SmartProxy build for?
The scripts define targets for chrome, chrome with service_worker, chrome-mv2, firefox, firefox-unlisted, firefox-android, edge and opera, each with a watch variant. Download links are given for Firefox on desktop and Android, the Chrome Web Store and the Edge Add-ons site, with none for Opera.
How do I build SmartProxy from source?
Node.js and a Firefox or Chrome browser are the stated requirements. The commands given are npm install followed by npm run build-ff:watch, with npm run build-ff, build-ch, build-ed and build-op for one-off builds per browser and further variants for unlisted, Android and mv2 targets.
What does SmartProxy use for IP geolocation?
IP2Location LITE, with the database deliberately left out of the repository because of licensing requirements. A checkout therefore cannot produce a complete build without that data arriving from somewhere outside the tree.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/salarcode-smartproxy)Star history chart for your README
[](https://hysenlabs.com/projects/salarcode-smartproxy)Was this useful?