PhishDetect is a library to help identify phishing pages

go get


Build Status Go Report Card GoDoc


NOTE: This project is experimental. It is not to be used yet, particularly with at-risk users.

PhishDetect is a library and a platform to detect potential phishing pages. It attempts doing so by identifying suspicious and malicious properties both in the domain names and URL provided, as well as in the HTML content of the page opened.

PhishDetect can take HTML strings as input, but it can also just be provided with an URL which will then be opened through a dedicated Docker container which automatically instruments a Google Chrome browser, whose behavior is monitored while navigating to the suspicious link.

Table of Contents


Install Docker Community Edition for Windows, Mac or Linux.

Download the Docker image from Docker Hub using:

$ docker pull phishdetect/phishdetect

Now you can download the PhishDetect library:

$ go get -u

For ease of versioning, you should consider using Go 1.11+ Modules in your own project.

Using PhishDetect as a library

Analyzing a link statically

You can then use it to analyze a URL or a domain like so:

package main

import (

func main() {
    // Instantiate an Analysis. The second argument is
    // an HTML string.
    a := phishdetect.NewAnalysis("", "")
    // Perform the analysis of the URL/domain.
    // Retrieve the name of the the impersonated brand.
    brand := a.Brands.GetBrand()

    // If the domain is recognized as whitelisted, this
    // will show as true, otherwise as false.
    // This is a total numeric value that is the sum of
    // all the score values of the warnings that were
    // matched during the analysis.
    // Print the brand. It will be an empty string if
    // no brand was identified.

    // Print all the matched warnings from the analysis.
    for _, warning := range a.Warnings {

Analyzing a link dynamically

If you want to analyze a URL by launching the dockerized Google Chrome:

package main

import (

func main() {
    url := ""
    // Instantiate a new Browser.
    // The first argument is the URL to analyze.
    // The second argument is the path to the file where to save the screenshot.
    // The third argument is a boolean value to enable or disable routing through Tor.
    b := phishdetect.NewBrowser(url, "/path/to/screen.png", false)
    // Run the browser.

    // Now we analyze the results.
    a := phishdetect.NewAnalysis(url, b.HTML)
    // Analyze the HTML string.
    brand := a.Brands.GetBrand()

    // In addition to the results explained in the previous example, we have
    // soma additional information provided by the browser execution.
    // FinalURL will show the last visited URL by the browser. This might differ
    // from the original URL if the browser was redirected.

    // Visits contains a list of all the URLs visited by the browser.
    // Normally 302 redirects or JavaScript redirects should appear (although in
    // the latter case, some might not appear if it took to long to load.)
    for _, visit := range b.Visits {

    // In addition to the URL analysis warnings, we should also have any matched
    // HTML analysis warnings.

For more information, please refer to the Godoc.

Using PhishDetect CLI

Firstly, make sure you have Go 1.11+ installed. We require Go 1.11 or later versions because of the native support for Go Modules, which we use to manage dependencies. If it isn't available for your operating system of choice, we recommend trying gvm.

Now you can either install PhishDetect's command-line interface by simply launching:

go get

Or build the binary from the source code. In order to do so, proceed cloning the Git repository:

$ git clone

Move to directory you just cloned and proceed with downloading the depedencies:

$ make deps

In order to build binaries for GNU/Linux:

$ make linux

In order to build binaries for Mac:

$ make darwin

In order to build binaries for Windows (this will require MingW to be installed):

$ go get
$ make windows

Once the compilation is completed, you will find the command-line interface, named phishdetect-cli, in the build/<linux/darwin/windows>/ folder

Launch phishdetect-cli -h to view the help message:

Usage of phishdetect-cli:
      --api-version string    Specify which Docker API version to use (default: 1.37) (default "1.37")
      --debug                 Enable debug logging (default: disabled)
      --safebrowsing string   Specify a file path containing your Google SafeBrowsing API key (default: disabled)
      --screen string         Specify the file path to store the screenshot (default: disabled)
      --tor                   Route connection through the Tor network (default: disabled)
      --url-only              Only perform URL analysis (default: disabled)

Specify a URL and the preferred options and wait for the results to appear:

$ build/linux/phishdetect-cli -screen /tmp/screen.png -tor http://[REDACTED].com/Login
INFO[0000] Analyzing URL http://[REDACTED].com/Login
INFO[0000] Using User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/62.0.3202.89 Safar$
INFO[0000] Using debug port: 9564
INFO[0000] Enabled route through the Tor network
INFO[0000] Started container with ID e43f6df4ab0fb8e29453df3ebaede0fe6a4bcbafa4fabaaa1da95573a28552ff
INFO[0000] Attempting to connect to debug port...
INFO[0001] Connection to debug port established!
INFO[0013] Saved screenshot at /tmp/screen.png
INFO[0013] Killed container with ID e43f6df4ab0fb8e29453df3ebaede0fe6a4bcbafa4fabaaa1da95573a28552ff
INFO[0013] Starting to analyze HTML...
INFO[0013] Matched password-input
INFO[0013] Matched suspicious-title
INFO[0014] Starting to analyze the URL...
INFO[0014] Matched suspicious-hostname
INFO[0014] Matched no-tls
INFO[0014] Visits:
INFO[0014]      - http://[REDACTED].com/Login
INFO[0014]      - http://[REDACTED].com/Login/
INFO[0014] Final URL: http://[REDACTED].com/Login/
INFO[0014] Whitelisted: false
INFO[0014] Final score: 90
INFO[0014] Brand: tutanota
INFO[0014] Warnings:
INFO[0014]      - The page contains a password input         name=password-input score=10
INFO[0014]      - The page has a suspicious title            name=suspicious-title score=30
INFO[0014]      - The domain contains suspicious words       name=suspicious-hostname score=30
INFO[0014]      - The website is not using a secure transport layer (HTTPS)  name=no-tls score=20


PhishDetect is released under GNU Affero General Public License 3.0 and is copyrighted to Claudio Guarnieri.