[GH-ISSUE #1589] Access internet #26639

Closed
opened 2026-04-22 03:02:15 -05:00 by GiteaMirror · 3 comments
Owner

Originally created by @PeachesMLG on GitHub (Dec 18, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/1589

Im customising my own model, using the steps in the ReadMe.

In this Modelfile I added a link to an faq with a bunch of information available, aswell as a github url in hopes it can search open/closed issues to awnser queries.

However it doesnt seem to be querying the url's like OpenAI GPT4 does

Is this currently possible?

Originally created by @PeachesMLG on GitHub (Dec 18, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/1589 Im customising my own model, using the steps in the ReadMe. In this Modelfile I added a link to an faq with a bunch of information available, aswell as a github url in hopes it can search open/closed issues to awnser queries. However it doesnt seem to be querying the url's like OpenAI GPT4 does Is this currently possible?
Author
Owner

@technovangelist commented on GitHub (Dec 19, 2023):

Hi @PeachesMLG

Thanks so much for submitting this issue. Models never have access to the outside world. Whenever you see a service that seems to give access to the outside world to a model, a developer has done some sort of search for relevant info, then sent that to the model for evaluation. And the same goes for us. There are a number of projects that you can find at the bottom of the readme on the repo that discuss RAG, or you can try to develop a solution yourself. Thanks again and let us know if you find anything else missing.

<!-- gh-comment-id:1862126404 --> @technovangelist commented on GitHub (Dec 19, 2023): Hi @PeachesMLG Thanks so much for submitting this issue. Models never have access to the outside world. Whenever you see a service that seems to give access to the outside world to a model, a developer has done some sort of search for relevant info, then sent that to the model for evaluation. And the same goes for us. There are a number of projects that you can find at the bottom of the readme on the repo that discuss RAG, or you can try to develop a solution yourself. Thanks again and let us know if you find anything else missing.
Author
Owner

@Danzelite commented on GitHub (Feb 5, 2024):

Hey idk if anyone will see this but due to the new release of the python/javascript library you can connect models to the web here is an example of a wiki-chatbot I made using the wiki-library for python and the ollama-python library:

import wikipedia, ollama

user_input = input("")

context = wikipedia.summary(user_input)

response = ollama.chat(model='qwen:0.5b', messages=[
{
'role': 'user',
'content': context,
},
])
print(response['message']['content'])

<!-- gh-comment-id:1927637155 --> @Danzelite commented on GitHub (Feb 5, 2024): Hey idk if anyone will see this but due to the new release of the python/javascript library you can connect models to the web here is an example of a wiki-chatbot I made using the wiki-library for python and the ollama-python library: import wikipedia, ollama user_input = input("") context = wikipedia.summary(user_input) response = ollama.chat(model='qwen:0.5b', messages=[ { 'role': 'user', 'content': context, }, ]) print(response['message']['content'])
Author
Owner

@areyoufeelingitnowmrkrebs commented on GitHub (Mar 5, 2025):

Hey idk if anyone will see this but due to the new release of the python/javascript library you can connect models to the web here is an example of a wiki-chatbot I made using the wiki-library for python and the ollama-python library:

import wikipedia, ollama

user_input = input("")

context = wikipedia.summary(user_input)

response = ollama.chat(model='qwen:0.5b', messages=[ { 'role': 'user', 'content': context, }, ]) print(response['message']['content'])

How do you use this to connect a model to the internet? I'm using phi4 on Arch.

<!-- gh-comment-id:2700269868 --> @areyoufeelingitnowmrkrebs commented on GitHub (Mar 5, 2025): > Hey idk if anyone will see this but due to the new release of the python/javascript library you can connect models to the web here is an example of a wiki-chatbot I made using the wiki-library for python and the ollama-python library: > > import wikipedia, ollama > > user_input = input("") > > context = wikipedia.summary(user_input) > > response = ollama.chat(model='qwen:0.5b', messages=[ { 'role': 'user', 'content': context, }, ]) print(response['message']['content']) How do you use this to connect a model to the internet? I'm using phi4 on Arch.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#26639