We are excited to announce that Qsirch 6.0 is now live!
This major update brings a smarter and more flexible search experience, while also removing the Premium license requirement β allowing more users to enjoy advanced features seamlessly.
Key Updates
On-Prem LLM + RAG Search (Beta)
Deploy on-prem LLMs (supports DeepSeek, Gemma, and more) on high-end QNAP NAS models with GPU support.
All data stays on your NAS β no cloud upload required β ensuring enhanced privacy and security.
Perfect for industries like finance, healthcare, and government that must comply with GDPR, HIPAA, and other data residency regulations.
RAG Multi-turn Conversations
AI can βremember what you just asked,β enabling contextual memory for a more natural, continuous conversation experience.
Ideal for step-by-step document analysis and in-depth Q&A β turning search into actionable insights.
seriously takes up CPU resources. The pause feature is great but it only gives at most a 1-week pause. Would like to keep it but going to can it with my CPU being run at 90%+ even after all the indexing and no one using the app.
I just went through a major QSirch indexing. It took the better part of a week. Depending on what you are indexing, how many media files you have, etc. it can take quite a while.
If you open Qsirch on the NAS, you can view what it is doing and how far it has to go from the tasks menu.
Once all the indexing is done, Qsirch settles down and everything is happy.
I have a TS-h1277AFX with Nvidia RTX 4000 Ada Generation (20GB) with QSirch 6.0.2 and LLM Core 1.0.0 and unfortunately no on-premise models appear for download. I have other QUTS NAS models that do show on-prem models for download, so it seems specific to this. The GPU is installed correctly and shows assigned to Container Station.
Anyone else had this issue and able to fix it? I opened a case awhile ago w/ QNAP Support and while they remoted in and duplicated the issue, theyβve been unable to provide any updates regarding any fix or workaround thus far.
Appreciate any help anyone can offer. And again Iβve got them working on another QNAP NAS (TS-h886 for example with a Geforce RTX 3050 6GB) so it seems to be specific to either the NAS model (TS-1277AFX) or the Nvidia GPU (RTX 4000 Ada Generation).
Looks like I have a quick answer to my own questionβ¦Looks like this was resolved with the LLM Core 1.0.1 qpkg version released today. On-prem LLM model download options now show up. Just sharing incase other users have this issue.