Skip to content
Switch branches/tags

dirsearch - Web path scanner

Build License Release Stars Tweet

Current Release: v0.4.1 (2020.12.8)


  • "dirsearch" is a mature command-line tool designed to brute force directories and files in webservers.

  • With 6 years of growth, dirsearch now has become the top web content scanner.

  • As a feature-rich tool, dirsearch gives users the opportunity to perform a complex web content discovering, with many vectors for the wordlist, high accuracy, impressive performance, advanced connection/request settings, modern brute-force techniques and nice output.

  • "dirsearch" is being actively developed by @maurosoria and @shelld3v

Installation & Usage

git clone
cd dirsearch
python3 -u <URL> -e <EXTENSIONS>

In case you want to run dirsearch anywhere:

git clone
cd dirsearch
pip3 install .
dirsearch -u <URL> -e <EXTENSIONS>
  • To can use SOCKS proxy, you need to install pips with requirements.txt: pip3 install -r requirements.txt

  • If you are using Windows and don't have git, you can install the ZIP file here. dirsearch also supports Docker

dirsearch requires python 3 or greater

About wordlists

Summary: Wordlist must be a text file, each line will be an endpoint. About extensions, unlike other tools, dirsearch doesn't append extensions to every word, if you don't use the -f flag. By default, only the %EXT% keyword in the wordlist will be replaced with extensions (-e <extensions>).


  • Each line in the wordlist will be processed as such, except when the special keyword %EXT% is used, it will generate one entry for each extension (-e | --extensions) passed as an argument.



Passing the extensions "asp" and "aspx" (-e asp,aspx) will generate the following dictionary:

  • For wordlists without %EXT% (like SecLists), you need to use the -f | --force-extensions switch to append extensions to every word in the wordlists, as well as the "/". And for entries in the wordlist that you do not want to force, you can add %NOFORCE% at the end of them so dirsearch won't append any extension.



Passing extensions "php" and "html" with the -f/--force-extensions flag (-f -e php,html) will generate the following dictionary:


To use multiple wordlists, you can seperate your wordlists with commas. Example: -w wordlist1.txt,wordlist2.txt


Usage: [-u|--url] target [-e|--extensions] extensions [options]

  --version             show program's version number and exit
  -h, --help            show this help message and exit

    -u URL, --url=URL   Target URL
    -l FILE, --url-list=FILE
                        Target URL list file
    --stdin             Target URL list from STDIN
    --cidr=CIDR         Target CIDR
    --raw=FILE          File contains the raw request (use `--scheme` flag to
                        set the scheme)
    -e EXTENSIONS, --extensions=EXTENSIONS
                        Extension list separated by commas (Example: php,asp)
    -X EXTENSIONS, --exclude-extensions=EXTENSIONS
                        Exclude extension list separated by commas (Example:
    -f, --force-extensions
                        Add extensions to every wordlist entry. By default
                        dirsearch only replaces the %EXT% keyword with

  Dictionary Settings:
    -w WORDLIST, --wordlists=WORDLIST
                        Customize wordlists (separated by commas)
                        Add custom prefixes to all wordlist entries (separated
                        by commas)
                        Add custom suffixes to all wordlist entries, ignore
                        directories (separated by commas)
    --only-selected     Remove paths have different extensions from selected
                        ones via `-e` (keep entries don't have extensions)
                        Remove extensions in all paths (Example: admin.php ->
    -U, --uppercase     Uppercase wordlist
    -L, --lowercase     Lowercase wordlist
    -C, --capital       Capital wordlist

  General Settings:
    -t THREADS, --threads=THREADS
                        Number of threads
    -r, --recursive     Brute-force recursively
                        Maximum recursion depth
                        Valid status codes to perform recursive scan, support
                        ranges (separated by commas)
    --subdirs=SUBDIRS   Scan sub-directories of the given URL[s] (separated by
                        Exclude the following subdirectories during recursive
                        scan (separated by commas)
    -i CODES, --include-status=CODES
                        Include status codes, separated by commas, support
                        ranges (Example: 200,300-399)
    -x CODES, --exclude-status=CODES
                        Exclude status codes, separated by commas, support
                        ranges (Example: 301,500-599)
                        Exclude responses by sizes, separated by commas
                        (Example: 123B,4KB)
                        Exclude responses by texts, separated by commas
                        (Example: 'Not found', 'Error')
                        Exclude responses by regexps, separated by commas
                        (Example: 'Not foun[a-z]{1}', '^Error$')
                        Exclude responses by redirect regexps or texts,
                        separated by commas (Example: '*')
                        Exclude responses by response content of this path
                        Skip target whenever hit one of these status codes,
                        separated by commas
    --minimal=LENGTH    Minimal response length
    --maximal=LENGTH    Maximal response length
    --max-time=SECONDS  Maximal runtime for the scan
    -q, --quiet-mode    Quiet mode
    --full-url          Full URLs in the output (enabled automatically in
                        quiet mode)
    --no-color          No colored output

  Request Settings:
    -m METHOD, --http-method=METHOD
                        HTTP method (default: GET)
    -d DATA, --data=DATA
                        HTTP request data
    -H HEADERS, --header=HEADERS
                        HTTP request header, support multiple flags (Example:
                        -H 'Referer:')
    --header-list=FILE  File contains HTTP request headers
    -F, --follow-redirects
                        Follow HTTP redirects
    --random-agent      Choose a random User-Agent for each request

  Connection Settings:
    --timeout=TIMEOUT   Connection timeout
    -s DELAY, --delay=DELAY
                        Delay between requests
    --proxy=PROXY       Proxy URL, support HTTP and SOCKS proxies (Example:
                        localhost:8080, socks5://localhost:8088)
    --proxy-list=FILE   File contains proxy servers
                        Proxy to replay with found paths
    --scheme=SCHEME     Default scheme (for raw request or if there is no
                        scheme in the URL)
                        Max requests per second
    --retries=RETRIES   Number of retries for failed requests
    -b, --request-by-hostname
                        By default dirsearch requests by IP for speed. This
                        will force dirsearch to request by hostname
    --ip=IP             Server IP address
    --exit-on-error     Exit whenever an error occurs

    -o FILE             Output file
    --format=FORMAT     Report format (Available: simple, plain, json, xml,
                        md, csv, html)

NOTE: You can change the dirsearch default configurations (default extensions, timeout, wordlist location, ...) by editing the default.conf file.

How to use

Dirsearch demo

Some examples for how to use dirsearch - those are the most common arguments. If you need all, just use the -h argument.

Simple usage

python3 -u https://target
python3 -e php,html,js -u https://target
python3 -e php,html,js -u https://target -w /path/to/wordlist

Recursive scan

By using the -r | --recursive argument, dirsearch will automatically brute-force the after of directories that it found.

python3 -e php,html,js -u https://target -r

You can set the max recursion depth with -R or --recursion-depth

python3 -e php,html,js -u https://target -r -R 3


The threads number (-t | --threads) reflects the number of separate brute force processes, that each process will perform path brute-forcing against the target. And so the bigger the threads number is, the more fast dirsearch runs. By default, the number of threads is 20, but you can increase it if you want to speed up the progress.

In spite of that, the speed is actually still uncontrollable since it depends a lot on the response time of the server. And as a warning, we advise you to keep the threads number not too big because of the impact from too much automation requests, this should be adjusted to fit the power of the system that you're scanning against.

python3 -e php,htm,js,bak,zip,tgz,txt -u https://target -t 30

Prefixes / Suffixes

  • --prefixes: Adding custom prefixes to all entries
python3 -e php -u https://target --prefixes .,admin,_

Base wordlist:


Generated with prefixes:

  • --suffixes: Adding custom suffixes to all entries
python3 -e php -u https://target --suffixes ~,/

Base wordlist:


Generated with suffixes:



Inside the db folder, there are several "blacklist files". Paths in those files will be filtered from the scan result if they have the same status as mentioned in the filename.

Example: If you add admin.php into db/403_blacklist.txt, whenever you do a scan that admin.php returns 403, it (admin.php) will be excluded.


Use -i | --include-status and -x | --exclude-status to select allowed and not allowed response status codes

python3 -e php,html,js -u https://target -i 200,204,400,403 -x 500,502,429

--exclude-sizes, --exclude-texts, --exclude-regexps, --exclude-redirects and --exclude-content are also supported for a more advanced filter

python3 -e php,html,js -u https://target --exclude-sizes 1B,243KB
python3 -e php,html,js -u https://target --exclude-texts "403 Forbidden"
python3 -e php,html,js -u https://target --exclude-regexps "^Error$"
python3 -e php,html,js -u https://target --exclude-content "admin.php"

Raw requests

dirsearch allows you to import the raw request from a file. The raw file content will be looked something like this:

GET /admin HTTP/1.1
Cache-Control: max-age=0
Accept: */*

Since there is no way for dirsearch to know what the URI scheme is (http or https), you need to set it using the --scheme flag. By default, the scheme is http, which is not popular in modern web servers now. That means, without setting up the scheme, you may brute-force with the wrong protocol, and will end up with false negatives.

Wordlist formats

Supported wordlist formats: uppercase, lowercase, capitalization







Exclude extensions

Use -X | --exclude-extensions with your exclude-extension list to remove all entries in the wordlist that have the given extensions

python3 -e asp,aspx -u https://target -X jsp

Base wordlist:




Scan sub-directories

From an URL, you can scan sub-directories with --subdirs.

python3 -e php,html,js -u https://target --subdirs admin/,folder/,/

A reverse version of this feature is --exclude-subdirs, which to prevent dirsearch from brute-forcing directories that should not be brute-forced when doing a recursive scan.

python3 -e php,html,js -u https://target --recursive -R 2 --exclude-subdirs "server-status/,%3f/"


Dirsearch supports SOCKS and HTTP proxy, with two options: a proxy server or a list of proxy servers.

python3 -e php,html,js -u https://target --proxy
python3 -e php,html,js -u https://target --proxy socks5://
python3 -e php,html,js -u https://target --proxylist proxyservers.txt


Dirsearch allows the user to save the output into a file. It supports several output formats like text or json, and we are keep updating for new formats

python3 -e php -l URLs.txt --format plain -o report.txt
python3 -e php -u https://target --format json -o target.json
python3 -e php -u https://target --format simple -o target.txt

Some others commands

python3 -e php,txt,zip -u https://target -w db/dicc.txt -H "X-Forwarded-Host:" -f
python3 -e php,txt,zip -u https://target -w db/dicc.txt -t 100 -m POST --data "username=admin"
python3 -e php,txt,zip -u https://target -w db/dicc.txt --random-agent --cookie "isAdmin=1"
python3 -e php,txt,zip -u https://target -w db/dicc.txt --format json -o target.json
python3 -e php,txt,zip -u https://target -w db/dicc.txt --minimal 1
python3 -e php,txt,zip -u https://target -w db/dicc.txt --header-list rate-limit-bypasses.txt
python3 -e php,txt,zip -u https://target -w db/dicc.txt -q --stop-on-error
python3 -e php,txt,zip -u https://target -w db/dicc.txt --full-url
python3 -u https://target -w db/dicc.txt --no-extension

There are more features and you will need to discover it by your self


  • To run dirsearch with a rate of requests per second, try -t <rate> -s 1
  • The server has a request limit? That's bad, but feel free to bypass it, by randomizing proxy with --proxy-list
  • Want to findout config files or backups? Try out --suffixes ~ and --prefixes .
  • For some endpoints that you do not want to force extensions, add %NOFORCE% at the end of them
  • Want to find only folders/directories? Combine --no-extension and --suffixes /!
  • The combination of --cidr, -F, -q and a low --timeout will reduce most of the noise + false negatives when brute-forcing with a CIDR

Support Docker

Install Docker Linux

Install Docker

curl -fsSL | bash

To use docker you need superuser power

Build Image dirsearch

To create image

docker build -t "dirsearch:v0.4.1" .

dirsearch is the name of the image and v0.4.1 is the version

Using dirsearch

For using

docker run -it --rm "dirsearch:v0.4.1" -u target -e php,html,js,zip


Copyright (C) Mauro Soria (

License: GNU General Public License, version 2


This tool is currently under development by @maurosoria and @shelld3v. We received a lot of help from many people around the world to improve this tool. Thanks so much to everyone who helped us!!

See for more information about who they are!

Want to join the team? Feel free to submit any pull request that you can. If you don't know how to code, you can support us by updating the wordlist or the documentation. Giving feedback or a new feature suggestion is also a good way to help us improve this tool