Cutting the Cord: Harnessing OpenAI's Whisper in an Offline Environment

Recently I presented a webinar on using ChatGPT for OSINT, where I used the entire one-hour window presenting. I encouraged attendees to ask questions in chat and said I would answer them on my blog. One of the questions was regarding another of OpenAI’s offerings, Whisper.

I recently blogged about Whisper, but the summary is that it’s an incredible audio-to-text transcription engine that can be installed on your local system and used 100% free.  The question was, does Whisper send the information to the cloud, or can it be done 100% offline? Offline would be fantastic from an OPSEC and privacy concerns point of view. I found one third-party blog post from when Whisper first launched, which mentioned that it could be used offline, but the official Whisper didn’t explicitly mention it.  In my reply, I said to expect to see a blog post on that soon, and this is that post.

I started off by using pip to install Whisper on a small NUC device running Ubuntu 22.04


Once that was done, I installed the opensource tool that specializes in audio and video files.


Here I'm confirming that Whisper was installed correctly. Notice the large number of languages supported. 


When asked to transcribe a video file I downloaded from YouTube, Whisper downloads its "medium" model by default. The file is only 461Mb in size and downloads quickly. Once downloaded, it correctly identifies the audio as being English and starts to transcribe it.

In order to test Whisper's offline capability, I disconnected the ethernet cable from the NUC and confirmed that the device was not connected to the internet. 

I then used Whisper to transcribe a video with English audio.


Here it's transcribing a video in Spanish.
Here I've asked Whisper to translate the Spanish audio to English, and it's doing it. All while still offline.



Whisper would be a fantastic tool no matter what, but the ability to transcribe video offline is a huge win for users who prioritize privacy and data security. Things like law enforcement transcribing interviews or intelligence analysts downloading and transcribing audio and videos on their focus area can utilize this tool and avoid a lot of the worries around OPSEC using third-party tools and services. 

 

Comments

Post a Comment

Popular posts from this blog

SANS Index How To Guide with Pictures

Image
I got some great advice recently on creating an index for SANS exams and I wanted to write a blog post to share it with others. I took the SANS FOR 508 Computer Forensics course in 2008. It was way over my head but I had a great time and learned a ton. A few months ago I finally decided to go for my GCFA certification. I had four year old material from a course that had been completely revamped and no index. I passed the exam with a score in the 80s but it was a grueling experience. I had to rush on the last part of the exam and never felt comfortable. A few months after my GCFA exam I got an opportunity to attend a SANS SEC 504 class. I really wanted to prepare for my GCIH exam the right way so while I was at the conference I asked several individuals how they prepared their index. Most people told me that their indexes were 8-10 pages. A lot of these people had more SANS certs than I have friends so their methods obviously worked for them. My class had a teaching assistant (also SANS
Read more

Introducing FaviconLocator: The Eazy Button to Searching by Favicon

Image
  Favicons (short for favorite icons) are the cute little pixelated images that appear next to the site name in web browser tabs, bookmarks, etc. In the image below we can see the iconic GitHub logo on their site and the KFC logo on a bucket of chicken on their site. Originally, favicons were designed to add a touch of professionalism and branding, but for anyone who is like me and has over a dozen tabs open at any time, favicons are the only thing displayed and how I navigate tabs. Most of us rely on favicons on a daily basis but many never think of them as a tool we can use in OSINT and CTI investigations. That’s what we’re doing to talk about here as well as introduce a new tool. In addition to branding and aiding in navigation between tabs, favicons can serve as unique identifiers for websites. These unique identifiers can help us: Trace the online presence of organizations and discover obscure digital assets Map the online infrastructure of potential threats Potentially de-ano
Read more

Automating Domain Squatting Detection with DNSTwist and Python

Image
  There’s a good chance that, at some point, you’ve received a spam email with a link that looked close to the name of a popular domain but was just a little off. Payapl.com instead of Paypal.com or similar. Domain squatting is a malicious activity where attackers register domain names similar to legitimate ones. Attackers may use these domains to deceive users into believing they are visiting a legitimate website, which can lead to phishing and other attacks. DNSTwist (available here: https://github.com/elceef/dnstwist ) is a popular tool that helps identify domain names that are similar to a given domain name. It generates a list of potential domain names by applying various fuzzing techniques to the given domain name and then checks if these domains are registered. Check out the list of Bank Of America copycat domains: DNSTwist is an amazing tool that should likely be a part of every organization's Cyber Threat Intelligence monitoring efforts, and I wanted to automate it
Read more