Communities

Writing
Writing
Codidact Meta
Codidact Meta
The Great Outdoors
The Great Outdoors
Photography & Video
Photography & Video
Scientific Speculation
Scientific Speculation
Cooking
Cooking
Electrical Engineering
Electrical Engineering
Judaism
Judaism
Languages & Linguistics
Languages & Linguistics
Software Development
Software Development
Mathematics
Mathematics
Christianity
Christianity
Code Golf
Code Golf
Music
Music
Physics
Physics
Linux Systems
Linux Systems
Power Users
Power Users
Tabletop RPGs
Tabletop RPGs
Community Proposals
Community Proposals
tag:snake search within a tag
answers:0 unanswered questions
user:xxxx search by author id
score:0.5 posts with 0.5+ score
"snake oil" exact phrase
votes:4 posts with 4+ votes
created:<1w created < 1 week ago
post_type:xxxx type of post
Search help
Notifications
Mark all as read See all your notifications »
Q&A

Welcome to the Power Users community on Codidact!

Power Users is a Q&A site for questions about the usage of computer software and hardware. We are still a small site and would like to grow, so please consider joining our community. We are looking forward to your questions and answers; they are the building blocks of a repository of knowledge we are building together.

Comments on When running a query with ChatGPT with search enabled, can I restrict the search to only certain domain(s)?

Post

When running a query with ChatGPT with search enabled, can I restrict the search to only certain domain(s)?

+0
−2

When running a query with ChatGPT with search enabled, can I restrict the search to only certain domain(s)?

For example, only search https://genai.stackexchange.com/


Crossposts:

History
Why does this post require attention from curators or moderators?
You might want to add some details to your flag.
Why should this post be closed?

1 comment thread

I don't think the answer to your cross-post on GenAI SE is correct. If the whole prompt is parsed by ... (7 comments)
I don't think the answer to your cross-post on GenAI SE is correct. If the whole prompt is parsed by ...
Andreas demands justice for humanity‭ wrote 7 days ago · edited 7 days ago

I don't think the answer to your cross-post on GenAI SE is correct. If the whole prompt is parsed by an LLM normally, then the following text that is generated is merely an approximation, a statistical likelihood, and not a guaranteed outcome. In other words, it's up to the LLM itself if it wants to respect your requirement. The LLM itself is, to the best of my knowledge, not designed to be able to do that, because it doesn't reason; it doesn't comply with strict rules, because the prompt isn't parsed to generate strict rules. It's read as tokens to generate new tokens.

In other words, the answer is "no" as long as there is no bypass for the prompt.

Would you like to comment on that? In case I'm wrong, I'd appreciate a correction.

Franck Dernoncourt‭ wrote 7 days ago · edited 7 days ago

Andreas demands justice for humanity‭ I think you're right (though still a useful answer if it works often enough, for some use cases). I've just checked the API too. OpenAI released a new API for web search in March 2025: https://platform.openai.com/docs/guides/tools-web-search. But it doesn't have any option yet to restrict the search to only certain domain(s).

Thanks. I decided to post an answer about that after receiving your confirmation and after reasoning a bit further about it. I'm not updated on the latest developments in LLMs, but I still haven't seen anything that's given me reason to reconsider my reservations against them. What that means, is that I wasn't comfortable posting an answer without refreshing and asserting my understanding of it.

Franck Dernoncourt‭ wrote 7 days ago · edited 7 days ago

Andreas demands justice for humanity‭ Thanks, this is difficult to stay up to date on LLM. E.g. the OpenAI search API just got released a month ago. This is one reason why I asked a few genAI questions.

Andreas demands justice for humanity‭ wrote 7 days ago · edited 7 days ago

That isn't as much of a concern. What I refer to, are the fundamentals of this form of generative AI. Addons and extensions don't change those bits; the design that incorporates certain attributes deep in the system. For example the hallucinations, or the inability to really reason. I'm unaware of any developments that have resolved these issues.

Andreas demands justice for humanity‭ The search API could've easily added domain restrictions.

It couldn't have influenced the training material.