Attachments such as images are in base64 and have to be extracted, you can use this python script to do that:

import os  
import email  
from email import policy  
import mimetypes  
import sys  


def extract_attachments(folder_path, output_root):  
    output_abs = os.path.abspath(output_root)  
    if not os.path.exists(output_root):  
        os.makedirs(output_root, exist_ok=True)  

    for sd in ["images", "videos", "documents"]:  
        os.makedirs(os.path.join(output_root, sd), exist_ok=True)  

    eml_files = []  
    for root, _, files in os.walk(folder_path):  
        if os.path.abspath(root).startswith(output_abs):  
            continue  
        for file in files:  
            if file.endswith(".eml"):  
                eml_files.append(os.path.join(root, file))  

    total_files = len(eml_files)  
    print(f"Found {total_files} .eml files to process.")  

    total_extracted = 0  
    for i, file_path in enumerate(eml_files, 1):  
        rel_path = os.path.relpath(file_path, folder_path)  
        print(f"[{i}/{total_files}] Processing: {rel_path}", end="\r", flush=True)  

        extracted_from_file = 0  
        try:  
            with open(file_path, "rb") as f:  
                msg = email.message_from_binary_file(f, policy=policy.default)  

                for part in msg.walk():  
                    if part.get_content_maintype() == "multipart":  
                        continue  

                    is_base64 = (  
                        part.get("Content-Transfer-Encoding", "").lower() == "base64"  
                    )  
                    is_media = part.get_content_maintype() in ["image", "video"]  

                    if is_base64 or is_media:  
                        filename = part.get_filename()  
                        if not filename:  
                            ext = (  
                                mimetypes.guess_extension(part.get_content_type())  
                                or ".bin"  
                            )  
                            filename = f"extracted_{hash(file_path)}_{id(part)}{ext}"  

                        maintype = part.get_content_maintype()  
                        target_dir = "documents"  
                        if maintype == "image":  
                            target_dir = "images"  
                        elif maintype == "video":  
                            target_dir = "videos"  

                        dest_path = os.path.join(output_root, target_dir, filename)  

                        base, extension = os.path.splitext(dest_path)  
                        counter = 1  
                        while os.path.exists(dest_path):  
                            dest_path = f"{base}_{counter}{extension}"  
                            counter += 1  

                        try:  
                            payload = part.get_payload(decode=True)  
                            if payload:  
                                with open(dest_path, "wb") as out_f:  
                                    out_f.write(payload)  
                                extracted_from_file += 1  
                                total_extracted += 1  
                        except Exception:  
                            pass  

        except Exception:  
            pass  

        if extracted_from_file > 0:  
            print(  
                f"[{i}/{total_files}] Extracted {extracted_from_file} from: {rel_path}"  
            )  

    print("\n" + "=" * 50)  
    return total_extracted  


if __name__ == "__main__":  
    SOURCE_DIR = "."  
    OUTPUT_DIR = "./extracted_attachments"  

    extracted_count = extract_attachments(SOURCE_DIR, OUTPUT_DIR)  
    print(f"Extraction complete. Total files extracted: {extracted_count}")  
    print(f"Files saved to: {os.path.abspath(OUTPUT_DIR)}")  

I’m so fucking over gatekeepers

Here is a direct link to the Kash Patel #leak in case anyone else was wanting it.

Pass: handala

https://link.storjshare.io/raw/jxoxwyp7qosgdwldereecudqpbva/poc/patelkpp%40gmail.com.zip

Source (Archive due to original being down): https://web.archive.org/web/20260327153947/https://handala-team.to/kash-patel-current-director-of-the-fbi-hacked/

Well so far so good

@aac1122 These corporate LLMs are trained to encourage user interaction so its no surprise you find yourself in agreement often, thats by design. Be careful how much you get into with these things, they can be very detrimental as well.

Always make sure you get feedback from humans to check that you aren’t just being told what the LLM thinks you want to hear.

Well it’s that time again, System keeps freezing up every once in a while for not much reason i can decern other then either bad RAM or bad SSD, replaced the SSD and it still froze so now here we are…

Other then having to figure out manual booting by commands in grub for the first time, homeserver’s migration to new SSD was great success!

The church has no place in how a state should be run. That's why we must separate religion from politics.

The cascadia #Meshcore is looking awesome.

Local #meshcore folk: check this out (and enable the Audio checkbox 🙂):

https://cascadiamesh.org/map/

my favourite part about open source is where you can’t contribute unless you subject yourself to using github and other corporate services.

I think about 90% of peoples problems with AI would go away if it wasnt being controlled by the absolute worst people for the absolute worst reasons, at the expense of everyone, while stealing everything….

LLMs are good tools, but anything these corporations do with it is the product of massive theft and exploitation and thats why so many people have such an averse reaction to it.

Take away the capitalism-demands-infinite-growth motive to LLMs and instead of data centers being build off a massive pyramid scheme of IOUs you’d have open community projects refining LLMs to make them more efficient and less demanding, but that’s not what drives hardware sales for Nvidia, so its not what the market wants.

Do you think Microsoft understands consent?

I’ve added Gitea and a handful of Nostr related services to GravityWell.xYz

fuck off.

imagine if windows required you to wait 24hrs before you could install programs outside of the MS store.

your phone is a computer, and this is an anti-competitive move that has the intentional side effect of incentivizing consumers to install apps from google instead of other markets.

“you can’t do this!”

“but i’ve always been able to do this!”

waits a year

“fine, but now you have to wait 24 hours because because because i said so!”

“oh! well in that case, i guess it’s not so bad!”

everyone that is happy about this is an easily manipulated buttbrain

@YoungBlood
Bro will claim they are so violent and unmanageable they might kill people at an airport but refuse to abolish them and keep giving them funding. What a worthless party full of worthless people.

after about 25 years of Midnight Commander I think I might finally be ready to make the switch to something new. Superfile is enough of an improvement that i’m willing to learn new hotkeys to use it.

Stop flirting in my 404 OMG /lh

I do think whenever this whole AI Bubble does pop we will end up with some very legit and useful applications for LLMs. They just wont be the sort of uses you’d be able to put directly into production without an amount of human oversight equal or greater to the amount of money you’d save by “eliminating” lower skilled workers. Like maybe you can hire less people maybe, but the people you do need will be much more expensive.

I’m not at all saying this makes LLMs worth paying for, in fact the only reason I even use Gemini over anything is because unlike similar tools it works for free and more over I was able to buy a google account on the darknet with a year of access to their ProAI for under $10 (none of which goes to google, its a free student account made by someone in someplace like Pakistan or Nigeria).