A simple TCP Connect port scanner developed in Python 3. This application leverages the use of Python's Standard Library asyncio framework to execute a number of TCP connections to an arbitrary number ports on target IP addresses, taking a maximum time equal to the connection timeout setting (defaults to 10 seconds) to return all results.

This application maintains no dependencies on third-party modules.

Version 2.0 rewrites the project as an installable, tested package: it fixes a crash that affected every Python 3.11+ interpreter under the prior release (a deprecated asyncio usage pattern that had gone unmaintained), adds a full test suite and CI, and ships on PyPI for the first time. The fix requires Python 3.12+; scan behavior and CLI flags are otherwise unchanged from prior versions. See CHANGELOG.md for the full release history.

user@host:~$ async-port-scanner scanme.nmap.org -p 20-25,53,80,111,135,139,443,3306,5900 --open
Starting Async Port Scanner at Wed Sep  9 11:19:33 2026
Scan report for scanme.nmap.org

[>] Results for scanme.nmap.org:
      PORT     STATE      SERVICE      REASON   
       22       open        ssh       SYN/ACK   
       80       open        http      SYN/ACK   

Async TCP Connect scan of 14 ports for scanme.nmap.org completed in 0.54 seconds

Install from PyPI, ideally with pipx so the async-port-scanner command is available globally in its own isolated environment:

user@host:~$ pipx install async-port-scanner

Or with pip:

user@host:~$ pip install async-port-scanner

user@host:~$ git clone https://github.com/EONRaider/Simple-Async-Port-Scanner.git
user@host:~$ cd Simple-Async-Port-Scanner
user@host:~/Simple-Async-Port-Scanner$ uv sync
user@host:~/Simple-Async-Port-Scanner$ uv run async-port-scanner example.com -p 80,443

usage: async-port-scanner [-h] -p PORTS [--timeout TIMEOUT] [--open]
                          [--version]
                          ADDRESSES

Simple asynchronous TCP Connect port scanner

positional arguments:
  ADDRESSES             A comma-separated sequence of IP addresses and/or domain names to scan, e.g., '45.33.32.156,65.61.137.117,testphp.vulnweb.com'.

options:
  -h, --help            show this help message and exit
  -p PORTS, --ports PORTS
                        A comma-separated sequence of port numbers and/or port ranges to scan on each target specified, e.g., '20-25,53,80,443'.
  --timeout TIMEOUT     Time to wait for a response from a target before closing a connection (defaults to 10.0 seconds).
  --open                Only show open ports in scan results.
  --version             show program's version number and exit

Usage examples:
1. async-port-scanner google.com -p 80,443
2. async-port-scanner 45.33.32.156,demo.testfire.net,18.192.172.30 -p 20-25,53,80,111,135,139,443,3306,5900

async-port-scanner is also runnable as a module with python -m async_port_scanner.

Thanks to Python's asyncio framework, a full sweep of the first 1000 TCP/IP ports on scanme.nmap.org typically completes in well under a second. Across 100 sequential runs of the exact command below (one run to completion before the next, with a 1.5-second pause between runs to avoid bursting the shared host), the median completion time was 0.64 seconds (mean 0.89s, range 0.57s-2.62s). Results were bimodal: ~77% of runs finished in under a second, while the rest clustered around 1.4-2.6s -- a pattern most likely explained by occasional TCP retransmission on the network path to the target, though this wasn't confirmed with a packet capture:

eonraider@havoc:~$ async-port-scanner scanme.nmap.org -p 1-1000 --open
Starting Async Port Scanner at Wed Sep  9 12:31:06 2026
Scan report for scanme.nmap.org

[>] Results for scanme.nmap.org:
      PORT     STATE      SERVICE      REASON   
       22       open        ssh       SYN/ACK   
       80       open        http      SYN/ACK   

Async TCP Connect scan of 1000 ports for scanme.nmap.org completed in 0.60 seconds

ADVISORY: For the sake of simplicity this application does not implement a maximum number of workers responsible for making each connection, instead spawning a new worker for every target socket (i.e. combination of target address and TCP port) until the process is complete. What this means in practice is that performing a scan of a significant number of ports on a single host will consequently trigger a great number of requests being sent almost simultaneously, potentially causing an involuntary situation analogous to that of a SYN-flood Denial-of-Service attack on hosts not able to handle the sudden spike in the number of requests they have to handle. For this particular reason, and in addition to the Legal Disclaimer section below, all users are advised by the developers to use caution when scanning live hosts.

cli.py's main() parses the target addresses and ports, builds an AsyncTCPScanner, and registers an OutputToScreen as its observer before the scan starts. AsyncTCPScanner._run() then drives one asyncio.TaskGroup task per target-port combination through _scan_target_port(), each attempting a TCP handshake bounded by --timeout and classifying the result — open, or closed with a reason — into a shared results dict. Once every task completes, the scanner notifies its registered observers and OutputToScreen prints the sorted results as a table. The full tour, including the concurrency tradeoff behind the ADVISORY above, is in ARCHITECTURE.md.

The test suite exercises this pipeline against local ephemeral ports rather than live network hosts, so it runs without any external dependency:

uv run pytest

Bug reports and pull requests are welcome. See CONTRIBUTING.md for the development setup and the lint/type-check/test suite that CI enforces on every push and pull request.

The use of code contained in this repository, either in part or in its totality, for engaging targets without prior mutual consent is illegal. It is the end user's responsibility to obey all applicable local, state and federal laws.

Developers assume no liability and are not responsible for misuses or damages caused by any code contained in this repository in any event that, accidentally or otherwise, it comes to be utilized by a threat agent or unauthorized entity as a means to compromise the security, privacy, confidentiality, integrity, and/or availability of systems and their associated resources by leveraging the exploitation of known or unknown vulnerabilities present in said systems, including, but not limited to, the implementation of security controls, human- or electronically-enabled.

The use of this code is only endorsed by the developers in those circumstances directly related to educational environments or authorized penetration testing engagements whose declared purpose is that of finding and mitigating vulnerabilities in systems, limiting their exposure to compromises and exploits employed by malicious agents as defined in their respective threat models.