Please consider making the Windows 32-bit versions Large Address Aware.

Tixati/Fopnu/DarkMX are all fairly memory intensive and can benefit from the (marginally) increased address-space on 32-bit Windows with almost no modification.

I've tested by running tixati.exe through ntcore's 4GB patcher (all it does is enable LAA in the PE header) on my laptop (which for testing is has a config just past the limit where an unpatched 32-bit version will start up, and patched it starts and runs fine.)

I know everyone should be running 64-bit if they at all can, but I wanted to post given how simple this is to implement and it's benefits.

So you want to add both old magnet archive and new magnet archive, and you are worried that the result might not fit into 2 GB of application memory? How much are you going to wait for a single search to finish then, half an hour? 50 requests per day feels like some limited access service for the rich. 8^)

Setting some flag means nothing by itself, you need to test that the application actually processes addresses from the extra area without bugs through all its functions. There might be some debug memory manager options to start allocations from the highest addresses, and let the applications crash early if something is wrong, but passing this test still doesn't guarantee that everything is correct in code.

In any case, better solutions are obvious, and you probably know them. Almost all x86 systems that support multiple gigabytes of physical memory can run in 64-bit mode, and use 64-bit operating system with unlimited address space. Exceptions are too slow for modern use (I don't say you can't do work with them, but that is often a sad experience). Maybe you want something special like ability to run unsigned drivers, etc., then maybe running Tixati somewhere else would be great. In addition to that, torrent/hash databases using actual big well known optimised database engines exist, and provide more features with greater speed. If I'm not mistaken, authors have used the same public data sets you can import, and query locally.