github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #914] Locally-hosted library #46957

New Issue
Open
opened 2026-04-28 02:16:05 -05:00 by GiteaMirror · 26 comments
GiteaMirror commented 2026-04-28 02:16:05 -05:00
Owner
Copy Link

Originally created by @JDRay42 on GitHub (Oct 26, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/914

Organizations that want to use Ollama in their enterprise will want some sort of control over the models that are available for use and where the trained models go when they get pushed. For instance, most implementers outside AI research won't necessarily want "uncensored" models on half the laptops in the company. And, if someone trains a model on corporate data, they will want to control where that model goes when someone pushes it, and won't want it made available to the general public.

So...

  1. As an enterprise implementer, I want to be able to host a local library.
  2. As an enterprise implementer, I want to be able to control the configuration of the Ollama instances on my network.
  3. As an enterprise implementer, I want to ensure that locally-developed libraries don't get pushed to a public repository.

I think that about covers it.

Originally created by @JDRay42 on GitHub (Oct 26, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/914 Organizations that want to use Ollama in their enterprise will want some sort of control over the models that are available for use and where the trained models go when they get pushed. For instance, most implementers outside AI research won't necessarily want "uncensored" models on half the laptops in the company. And, if someone trains a model on corporate data, they will want to control where that model goes when someone pushes it, and won't want it made available to the general public. So... 1. As an enterprise implementer, I want to be able to host a local library. 2. As an enterprise implementer, I want to be able to control the configuration of the Ollama instances on my network. 3. As an enterprise implementer, I want to ensure that locally-developed libraries don't get pushed to a public repository. I think that about covers it.
GiteaMirror added the question label 2026-04-28 02:16:05 -05:00
GiteaMirror commented 2026-04-28 02:16:11 -05:00
Author
Owner
Copy Link

@mchiang0610 commented on GitHub (Oct 26, 2023):

Hey @JDRay42 thank you for creating this issue. Are there particular organizations you are working with who require these features right now? I can understand if it's sensitive - my email is michael@ollama.ai

<!-- gh-comment-id:1781391975 --> @mchiang0610 commented on GitHub (Oct 26, 2023): Hey @JDRay42 thank you for creating this issue. Are there particular organizations you are working with who require these features right now? I can understand if it's sensitive - my email is michael@ollama.ai
GiteaMirror commented 2026-04-28 02:16:12 -05:00
Author
Owner
Copy Link

@JDRay42 commented on GitHub (Oct 26, 2023):

I just started work for an organization, and they're keen to understand what AI can do for them. Due to the nature of the organization, they're concerned about using tools like ChatGPT and whatnot because of data security. I'm putting together a series of presentations, and will suggest that we implement a local, server-hosted tool like Ollama, as an entry into the space. From there, making the app available to people for local (laptop) install would be an advancement, but not required. I'll guide them toward RAG workflows and LoRA adapter development, though.

I'll send you some email.

<!-- gh-comment-id:1781420467 --> @JDRay42 commented on GitHub (Oct 26, 2023): I just started work for an organization, and they're keen to understand what AI can do for them. Due to the nature of the organization, they're concerned about using tools like ChatGPT and whatnot because of data security. I'm putting together a series of presentations, and will suggest that we implement a local, server-hosted tool like Ollama, as an entry into the space. From there, making the app available to people for local (laptop) install would be an advancement, but not required. I'll guide them toward RAG workflows and LoRA adapter development, though. I'll send you some email.
GiteaMirror commented 2026-04-28 02:16:14 -05:00
Author
Owner
Copy Link

@taha-yassine commented on GitHub (Nov 7, 2023):

Hi, I'm interested in this use case as well. Having a standalone server in charge of hosting and serving the models in the style of Docker registries would be a real game changer and would bring the project to the next level IMHO.

<!-- gh-comment-id:1799025217 --> @taha-yassine commented on GitHub (Nov 7, 2023): Hi, I'm interested in this use case as well. Having a standalone server in charge of hosting and serving the models in the style of Docker registries would be a real game changer and would bring the project to the next level IMHO.
GiteaMirror commented 2026-04-28 02:16:18 -05:00
Author
Owner
Copy Link

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

We are also interested in this for our own initiative. We hope we could use Artifactory as a source for Ollama models.

<!-- gh-comment-id:1863274084 --> @ryancurrah commented on GitHub (Dec 19, 2023): We are also interested in this for our own initiative. We hope we could use Artifactory as a source for Ollama models.
GiteaMirror commented 2026-04-28 02:16:22 -05:00
Author
Owner
Copy Link

@kamuridesu commented on GitHub (Jan 16, 2024):

No updates into this?
I think it would be grate to integrate with the model registry from OpenDataHub: https://github.com/opendatahub-io/model-registry

<!-- gh-comment-id:1894559101 --> @kamuridesu commented on GitHub (Jan 16, 2024): No updates into this? I think it would be grate to integrate with the model registry from OpenDataHub: https://github.com/opendatahub-io/model-registry
GiteaMirror commented 2026-04-28 02:16:28 -05:00
Author
Owner
Copy Link

@TheComputerDan commented on GitHub (Jan 26, 2024):

We are also interested in this for our own initiative. We hope we could use Artifactory as a source for Ollama models.

I think this would be great. Support for Binary Object Repositories such as Artifactory and Nexus would open the door to many organizations who have additional security controls, or may want to develop and publish privately as opposed to a cloud only service.

<!-- gh-comment-id:1911199180 --> @TheComputerDan commented on GitHub (Jan 26, 2024): > We are also interested in this for our own initiative. We hope we could use Artifactory as a source for Ollama models. I think this would be great. Support for Binary Object Repositories such as Artifactory and Nexus would open the door to many organizations who have additional security controls, or may want to develop and publish privately as opposed to a cloud only service.
GiteaMirror commented 2026-04-28 02:16:31 -05:00
Author
Owner
Copy Link

@ekk1 commented on GitHub (Feb 6, 2024):

Also interested in this, we have many servers that needs to pull models, it will be a lot faster if we can pull from a local registry. Currently we need to distribute the model files using other tools, it would be very helpful if ollama has this builtin.

<!-- gh-comment-id:1928869130 --> @ekk1 commented on GitHub (Feb 6, 2024): Also interested in this, we have many servers that needs to pull models, it will be a lot faster if we can pull from a local registry. Currently we need to distribute the model files using other tools, it would be very helpful if ollama has this builtin.
GiteaMirror commented 2026-04-28 02:16:32 -05:00
Author
Owner
Copy Link

@chboerner commented on GitHub (Feb 16, 2024):

+1 - would be interested in this feature as well

<!-- gh-comment-id:1949268064 --> @chboerner commented on GitHub (Feb 16, 2024): +1 - would be interested in this feature as well
GiteaMirror commented 2026-04-28 02:16:34 -05:00
Author
Owner
Copy Link

@jamesbrink commented on GitHub (Feb 20, 2024):

To run your own registry server simply use https://distribution.github.io/distribution/

Quick example of running locally:

docker run -d -p 5001:5000 --name registry registry:2
ollama cp llama2 localhost:5001/jamesbrink/llama2
ollama push localhost:5001/jamesbrink/llama2 --insecure

You can quickly verify the image using curl:

curl -s -H "Accept: application/vnd.docker.distribution.manifest.v2+json" -H "Content-Type: application/json" http://localhost:5001/v2/jamesbrink/llama2/manifests/latest | jq '.'
{
  "schemaVersion": 2,
  "mediaType": "application/vnd.docker.distribution.manifest.v2+json",
  "config": {
    "mediaType": "application/vnd.docker.container.image.v1+json",
    "digest": "sha256:42ba7f8a01ddb4fa59908edd37d981d3baa8d8efea0e222b027f29f7bcae21f9",
    "size": 557
  },
  "layers": [
    {
      "mediaType": "application/vnd.ollama.image.model",
      "digest": "sha256:8934d96d3f08982e95922b2b7a2c626a1fe873d7c3b06e8e56d7bc0a1fef9246",
      "size": 3826781184
    },
    {
      "mediaType": "application/vnd.ollama.image.license",
      "digest": "sha256:8c17c2ebb0ea011be9981cc3922db8ca8fa61e828c5d3f44cb6ae342bf80460b",
      "size": 7020
    },
    {
      "mediaType": "application/vnd.ollama.image.license",
      "digest": "sha256:7c23fb36d80141c4ab8cdbb61ee4790102ebd2bf7aeff414453177d4f2110e5d",
      "size": 4766
    },
    {
      "mediaType": "application/vnd.ollama.image.template",
      "digest": "sha256:2e0493f67d0c8c9c68a8aeacdf6a38a2151cb3c4c1d42accf296e19810527988",
      "size": 59
    },
    {
      "mediaType": "application/vnd.ollama.image.params",
      "digest": "sha256:fa304d6750612c207b8705aca35391761f29492534e90b30575e4980d6ca82f6",
      "size": 91
    }
  ]
}
<!-- gh-comment-id:1953482174 --> @jamesbrink commented on GitHub (Feb 20, 2024): To run your own registry server simply use https://distribution.github.io/distribution/ Quick example of running locally: ```shell docker run -d -p 5001:5000 --name registry registry:2 ollama cp llama2 localhost:5001/jamesbrink/llama2 ollama push localhost:5001/jamesbrink/llama2 --insecure ``` You can quickly verify the image using curl: ```shell curl -s -H "Accept: application/vnd.docker.distribution.manifest.v2+json" -H "Content-Type: application/json" http://localhost:5001/v2/jamesbrink/llama2/manifests/latest | jq '.' { "schemaVersion": 2, "mediaType": "application/vnd.docker.distribution.manifest.v2+json", "config": { "mediaType": "application/vnd.docker.container.image.v1+json", "digest": "sha256:42ba7f8a01ddb4fa59908edd37d981d3baa8d8efea0e222b027f29f7bcae21f9", "size": 557 }, "layers": [ { "mediaType": "application/vnd.ollama.image.model", "digest": "sha256:8934d96d3f08982e95922b2b7a2c626a1fe873d7c3b06e8e56d7bc0a1fef9246", "size": 3826781184 }, { "mediaType": "application/vnd.ollama.image.license", "digest": "sha256:8c17c2ebb0ea011be9981cc3922db8ca8fa61e828c5d3f44cb6ae342bf80460b", "size": 7020 }, { "mediaType": "application/vnd.ollama.image.license", "digest": "sha256:7c23fb36d80141c4ab8cdbb61ee4790102ebd2bf7aeff414453177d4f2110e5d", "size": 4766 }, { "mediaType": "application/vnd.ollama.image.template", "digest": "sha256:2e0493f67d0c8c9c68a8aeacdf6a38a2151cb3c4c1d42accf296e19810527988", "size": 59 }, { "mediaType": "application/vnd.ollama.image.params", "digest": "sha256:fa304d6750612c207b8705aca35391761f29492534e90b30575e4980d6ca82f6", "size": 91 } ] } ```
GiteaMirror commented 2026-04-28 02:16:37 -05:00
Author
Owner
Copy Link

@easp commented on GitHub (Feb 20, 2024):

@jamesbrink Thanks for sharing this. I've long suspected it was possible due to some clues in the CLI options and a few statements made by maintainers in the past. I didn't know anything about docker registries and I didn't have any need for it though, so I never bothered figuring it out.

<!-- gh-comment-id:1954593379 --> @easp commented on GitHub (Feb 20, 2024): @jamesbrink Thanks for sharing this. I've long suspected it was possible due to some clues in the CLI options and a few statements made by maintainers in the past. I didn't know anything about docker registries and I didn't have any need for it though, so I never bothered figuring it out.
GiteaMirror commented 2026-04-28 02:16:39 -05:00
Author
Owner
Copy Link

@magnusheino commented on GitHub (Mar 5, 2024):

Looking at that exemple of running a registry, it looks like it should be pretty easy to get setup a Artifactory remote proxy... https://jfrog.com/help/r/jfrog-artifactory-documentation/remote-docker-repositories

But I don't get it to work. Is there anything in the implementation/protocol that I should be aware of that is different from docker?

<!-- gh-comment-id:1978588128 --> @magnusheino commented on GitHub (Mar 5, 2024): Looking at that exemple of running a registry, it looks like it should be pretty easy to get setup a Artifactory remote proxy... https://jfrog.com/help/r/jfrog-artifactory-documentation/remote-docker-repositories But I don't get it to work. Is there anything in the implementation/protocol that I should be aware of that is different from docker?
GiteaMirror commented 2026-04-28 02:16:42 -05:00
Author
Owner
Copy Link

@ccll commented on GitHub (Mar 15, 2024):

Same here, we're a small Chinese startup that focus on cybersecurity solutions, our code cannot upload to public AI models like ChatGPT or CodeWhisperer due to client policies, so we've setup multiple ollama deployments inside the company for internal use, since the network connection is unstable due to the existence of GFW, we desperately need to cache those models to speedup downloading.

For docker containers we use SonaType Nexus as a pull through cache/proxy, it works very well and saved us a lot of time and effort, would this be possible with ollama?

<!-- gh-comment-id:1998824040 --> @ccll commented on GitHub (Mar 15, 2024): Same here, we're a small Chinese startup that focus on cybersecurity solutions, our code cannot upload to public AI models like ChatGPT or CodeWhisperer due to client policies, so we've setup multiple ollama deployments inside the company for internal use, since the network connection is unstable due to the existence of GFW, we desperately need to cache those models to speedup downloading. For docker containers we use SonaType Nexus as a pull through cache/proxy, it works very well and saved us a lot of time and effort, would this be possible with ollama?
GiteaMirror commented 2026-04-28 02:16:43 -05:00
Author
Owner
Copy Link

@jamesbrink commented on GitHub (Apr 10, 2024):

Regarding this comment: https://github.com/ollama/ollama/issues/2388#issuecomment-1989307410
@bmizerany is there any chance you guys could open source some of your backend? I am sure it's probably not hard to implement something like this with Distribution Registry, but I have only skimmed the code so I truly don't know. This starts to get into areas I am truly not that familiar with, but I am optimistic you guys can keep it compatible or at least open source the back end eventually.

<!-- gh-comment-id:2046286021 --> @jamesbrink commented on GitHub (Apr 10, 2024): Regarding this comment: https://github.com/ollama/ollama/issues/2388#issuecomment-1989307410 @bmizerany is there any chance you guys could open source some of your backend? I am sure it's probably not hard to implement something like this with Distribution Registry, but I have only skimmed the code so I truly don't know. This starts to get into areas I am truly not that familiar with, but I am optimistic you guys can keep it compatible or at least open source the back end eventually.
GiteaMirror commented 2026-04-28 02:16:45 -05:00
Author
Owner
Copy Link

@RWayne93 commented on GitHub (May 4, 2024):

+1 for this issue as well. I have successfully hosted my own models in a registry following @jamesbrink method however now i have been trying to do this with a protected registry using a self signed certificate however i keep encountering x509 errors.

<!-- gh-comment-id:2094288422 --> @RWayne93 commented on GitHub (May 4, 2024): +1 for this issue as well. I have successfully hosted my own models in a registry following @jamesbrink method however now i have been trying to do this with a protected registry using a self signed certificate however i keep encountering x509 errors.
GiteaMirror commented 2026-04-28 02:16:46 -05:00
Author
Owner
Copy Link

@ifelsefi commented on GitHub (May 30, 2024):

Looking at that exemple of running a registry, it looks like it should be pretty easy to get setup a Artifactory remote proxy... https://jfrog.com/help/r/jfrog-artifactory-documentation/remote-docker-repositories

But I don't get it to work. Is there anything in the implementation/protocol that I should be aware of that is different from docker?

I tried to get this working but ollama seems to prepend the public registry URL to our Artifactory URL:

llama3:

user@gpunode[~]$ ollama pull registry-ollama-ai.fqdn/llama3
pulling manifest
Error: pull model manifest: Get "https://registry.ollama.ai/v2/registry-ollama-ai.fqdn/llama3/manifests/latest": dial tcp 172.67.182.229:443: i/o timeout

starcoder:

user@gpu-node[~]$ ollama pull registry-ollama-ai.fqdn/starcoder
pulling manifest
Error: pull model manifest: Get "https://registry.ollama.ai/v2/registry-ollama-ai.fqdn/starcoder/manifests/latest": dial tcp 172.67.182.229:443: i/o timeout

Running:

user@gpu-node[~]$ ollama --version
ollama version is 0.1.39

Any idea now to fix?

ollama does not work with Squid proxy server either. These tools are required for enterprise adoption.

<!-- gh-comment-id:2140877092 --> @ifelsefi commented on GitHub (May 30, 2024): > Looking at that exemple of running a registry, it looks like it should be pretty easy to get setup a Artifactory remote proxy... https://jfrog.com/help/r/jfrog-artifactory-documentation/remote-docker-repositories > > But I don't get it to work. Is there anything in the implementation/protocol that I should be aware of that is different from docker? I tried to get this working but `ollama` seems to prepend the public registry URL to our Artifactory URL: `llama3`: ``` user@gpunode[~]$ ollama pull registry-ollama-ai.fqdn/llama3 pulling manifest Error: pull model manifest: Get "https://registry.ollama.ai/v2/registry-ollama-ai.fqdn/llama3/manifests/latest": dial tcp 172.67.182.229:443: i/o timeout ``` `starcoder`: ``` user@gpu-node[~]$ ollama pull registry-ollama-ai.fqdn/starcoder pulling manifest Error: pull model manifest: Get "https://registry.ollama.ai/v2/registry-ollama-ai.fqdn/starcoder/manifests/latest": dial tcp 172.67.182.229:443: i/o timeout ``` Running: ``` user@gpu-node[~]$ ollama --version ollama version is 0.1.39 ``` Any idea now to fix? ollama does not work with Squid proxy server either. These tools are required for enterprise adoption.
GiteaMirror commented 2026-04-28 02:16:47 -05:00
Author
Owner
Copy Link

@jorander commented on GitHub (Jun 7, 2024):

This is the code in Ollama setting up the url to use when pulling a model: https://github.com/ollama/ollama/blob/main/server/modelpath.go#L50-L62
It uses a set of default values defined here: https://github.com/ollama/ollama/blob/main/server/modelpath.go#L21-L26

If it finds to few parts when dividing the model name by / it uses the defaults to fill in the gaps. If it finds to many it uses the default entirely.

This is the problem I see when trying to use Ollama with Artifactory as a mirror for registry.ollama.ai. When setting up Artifactory as a mirror I need to set up an Artifactory remote repository with an arbitrary name (say "ollama-models-remote") of type Docker and with the url set to https://registry.ollama.ai

For Artifactory to call Ollama model registry it needs to use https://registry.ollama.ai/v2/library/... . The v2 is added by Artifactory itself after the configured url. library needs to be supplied by me, or by my local Ollama server to be more precise, when calling my Artifactory repo. This is common practice as this is the namespace within the registry.ollama.ai where the model I'm interested in is located.

To download a model through my local Artifactory repo I would need to name it my.artifacory.host/ollama-models-remote/library/<model>:<tag>. This should cause the local Ollama server to call https://my.artifacory.host/v2/ollama-models-remote/library/..., which would cause Artifactory to substitute https://my.artifacory.host/v2/ollama-models-remote/ for https://registry.ollama.ai/v2/ and then append library/.... But since the code in Ollama used to split the model name only handles up to three parts divided by / and my supplied name contains four, it goes with the default instead.

I think only a small change, adding a forth case where part 1 and 2 are combined with a / to build the namespace (mp.Namespace = parts[1] + "/" + parts[2]), would resolve this issue.

<!-- gh-comment-id:2155062721 --> @jorander commented on GitHub (Jun 7, 2024): This is the code in Ollama setting up the url to use when pulling a model: https://github.com/ollama/ollama/blob/main/server/modelpath.go#L50-L62 It uses a set of default values defined here: https://github.com/ollama/ollama/blob/main/server/modelpath.go#L21-L26 If it finds to few parts when dividing the model name by `/` it uses the defaults to fill in the gaps. If it finds to many it uses the default entirely. This is the problem I see when trying to use Ollama with Artifactory as a mirror for `registry.ollama.ai`. When setting up Artifactory as a mirror I need to set up an Artifactory remote repository with an arbitrary name (say "ollama-models-remote") of type Docker and with the url set to `https://registry.ollama.ai` For Artifactory to call Ollama model registry it needs to use `https://registry.ollama.ai/v2/library/...` . The `v2` is added by Artifactory itself after the configured url. `library` needs to be supplied by me, or by my local Ollama server to be more precise, when calling my Artifactory repo. This is common practice as this is the namespace within the `registry.ollama.ai` where the model I'm interested in is located. To download a model through my local Artifactory repo I would need to name it `my.artifacory.host/ollama-models-remote/library/<model>:<tag>`. This should cause the local Ollama server to call `https://my.artifacory.host/v2/ollama-models-remote/library/...`, which would cause Artifactory to substitute `https://my.artifacory.host/v2/ollama-models-remote/` for `https://registry.ollama.ai/v2/` and then append `library/...`. But since the code in Ollama used to split the model name only handles up to three parts divided by `/` and my supplied name contains four, it goes with the default instead. I think only a small change, adding a forth case where part 1 and 2 are combined with a `/ ` to build the namespace (`mp.Namespace = parts[1] + "/" + parts[2]`), would resolve this issue.
GiteaMirror commented 2026-04-28 02:16:50 -05:00
Author
Owner
Copy Link

@stonezdj commented on GitHub (Jun 21, 2024):

Same here, we're a small Chinese startup that focus on cybersecurity solutions, our code cannot upload to public AI models like ChatGPT or CodeWhisperer due to client policies, so we've setup multiple ollama deployments inside the company for internal use, since the network connection is unstable due to the existence of GFW, we desperately need to cache those models to speedup downloading.

For docker containers we use SonaType Nexus as a pull through cache/proxy, it works very well and saved us a lot of time and effort, would this be possible with ollama?

You could try the Harbor registry, there is a dev build you could try:
https://github.com/stonezdj/harbor/releases/v2.11.0-ollama

<!-- gh-comment-id:2181991058 --> @stonezdj commented on GitHub (Jun 21, 2024): > Same here, we're a small Chinese startup that focus on cybersecurity solutions, our code cannot upload to public AI models like ChatGPT or CodeWhisperer due to client policies, so we've setup multiple ollama deployments inside the company for internal use, since the network connection is unstable due to the existence of GFW, we desperately need to cache those models to speedup downloading. > > For docker containers we use SonaType Nexus as a pull through cache/proxy, it works very well and saved us a lot of time and effort, would this be possible with ollama? You could try the Harbor registry, there is a dev build you could try: https://github.com/stonezdj/harbor/releases/v2.11.0-ollama
GiteaMirror commented 2026-04-28 02:16:52 -05:00
Author
Owner
Copy Link

@simonfrey commented on GitHub (Jul 2, 2024):

As ollama is somewhat docker, but also somewhat not, I felt we need a custom solution. Hence I am building an ollama registry pull through cache (https://github.com/beans-bytes/ollama-registry-pull-through-proxy)

Running it is quite straight forward via docker:

  1. Start the cache docker run -p 9200:9200 -v ./cache_dir_docker:/pull-through-cache ghcr.io/beans-bytes/ollama-registry-pull-through-proxy:latest
  2. Pull/run against it ollama pull --insecure http://localhost:9200/library/<image>:<tag>

WDYT?

<!-- gh-comment-id:2202709638 --> @simonfrey commented on GitHub (Jul 2, 2024): As ollama is somewhat docker, but also somewhat not, I felt we need a custom solution. Hence I am building an ollama registry pull through cache (https://github.com/beans-bytes/ollama-registry-pull-through-proxy) Running it is quite straight forward via docker: 1. Start the cache `docker run -p 9200:9200 -v ./cache_dir_docker:/pull-through-cache ghcr.io/beans-bytes/ollama-registry-pull-through-proxy:latest` 2. Pull/run against it `ollama pull --insecure http://localhost:9200/library/<image>:<tag>` WDYT?
GiteaMirror commented 2026-04-28 02:16:55 -05:00
Author
Owner
Copy Link

@wbuntu commented on GitHub (Sep 12, 2024):

The ollama registry seems a little diffrence from docker reigstry，but the pull logic is simple.
I just build a pull-only local registry with caddy and ollama：https://github.com/wbuntu/images/tree/master/ollama-registry
Also it can be used as an offline model reigstry.

<!-- gh-comment-id:2345514451 --> @wbuntu commented on GitHub (Sep 12, 2024): The ollama registry seems a little diffrence from docker reigstry，but the pull logic is simple. I just build a pull-only local registry with caddy and ollama：https://github.com/wbuntu/images/tree/master/ollama-registry Also it can be used as an offline model reigstry.
GiteaMirror commented 2026-04-28 02:16:56 -05:00
Author
Owner
Copy Link

@cniweb commented on GitHub (Nov 27, 2024):

We are also interested in this for our own initiative. We hope we could use Artifactory as a source for Ollama models.

I think this would be great. Support for Binary Object Repositories such as Artifactory and Nexus would open the door to many organizations who have additional security controls, or may want to develop and publish privately as opposed to a cloud only service.

Yes, that's exactly what we need as an organization that is regulated and requires content security.
We use Nexus Repository as a proxy to cache artifacts internally.

<!-- gh-comment-id:2503063567 --> @cniweb commented on GitHub (Nov 27, 2024): > > We are also interested in this for our own initiative. We hope we could use Artifactory as a source for Ollama models. > > I think this would be great. Support for Binary Object Repositories such as Artifactory and Nexus would open the door to many organizations who have additional security controls, or may want to develop and publish privately as opposed to a cloud only service. Yes, that's exactly what we need as an organization that is regulated and requires content security. We use Nexus Repository as a proxy to cache artifacts internally.
GiteaMirror commented 2026-04-28 02:16:57 -05:00
Author
Owner
Copy Link

@cniweb commented on GitHub (Jan 9, 2025):

Same issue as #5298

<!-- gh-comment-id:2579356539 --> @cniweb commented on GitHub (Jan 9, 2025): Same issue as #5298
GiteaMirror commented 2026-04-28 02:16:58 -05:00
Author
Owner
Copy Link

@xuxinping commented on GitHub (Feb 2, 2025):

The artifact repository, Folib, supports the private management of Ollama models. For specific inquiries, please visit www.bocloud.com.cn.

Image

<!-- gh-comment-id:2629428322 --> @xuxinping commented on GitHub (Feb 2, 2025): The artifact repository, Folib, supports the private management of Ollama models. For specific inquiries, please visit [www.bocloud.com.cn](https://www.bocloud.com.cn/). ![Image](https://github.com/user-attachments/assets/74951e3f-e609-4e17-b6d0-d66a22d74bef)
GiteaMirror commented 2026-04-28 02:16:58 -05:00
Author
Owner
Copy Link

@thawkins commented on GitHub (Apr 23, 2025):

We are also interested in this for our own initiative. We hope we could use Artifactory as a source for Ollama models.

I think this would be great. Support for Binary Object Repositories such as Artifactory and Nexus would open the door to many organizations who have additional security controls, or may want to develop and publish privately as opposed to a cloud only service.

Yes, i work for a large global fintech with 18k developers, we are obviously spining up a lot ai projects, and given we are a fintech most of our data us very sensative. We would like to use ollama effectivly to "mock" public hosted apis via the OpenAI api emulation. We would also like to be able to host internal repos of models that can be sourced by pull through from the artifactory system acting as a caching proxy.

<!-- gh-comment-id:2823990876 --> @thawkins commented on GitHub (Apr 23, 2025): > > We are also interested in this for our own initiative. We hope we could use Artifactory as a source for Ollama models. > > I think this would be great. Support for Binary Object Repositories such as Artifactory and Nexus would open the door to many organizations who have additional security controls, or may want to develop and publish privately as opposed to a cloud only service. Yes, i work for a large global fintech with 18k developers, we are obviously spining up a lot ai projects, and given we are a fintech most of our data us very sensative. We would like to use ollama effectivly to "mock" public hosted apis via the OpenAI api emulation. We would also like to be able to host internal repos of models that can be sourced by pull through from the artifactory system acting as a caching proxy.
GiteaMirror commented 2026-04-28 02:16:59 -05:00
Author
Owner
Copy Link

@jorander commented on GitHub (Apr 23, 2025):

@thawkins I work in a much smaller organisation, but with same type of requirements when it comes to keeping data private. In this Jfokus conference talk from earlier this year we talk (among other things) about how we've set up Ollama to work in our internal IT-environment.

https://www.youtube.com/watch?v=CPF4UZatFZQ

<!-- gh-comment-id:2824026760 --> @jorander commented on GitHub (Apr 23, 2025): @thawkins I work in a much smaller organisation, but with same type of requirements when it comes to keeping data private. In this Jfokus conference talk from earlier this year we talk (among other things) about how we've set up Ollama to work in our internal IT-environment. https://www.youtube.com/watch?v=CPF4UZatFZQ
GiteaMirror commented 2026-04-28 02:17:00 -05:00
Author
Owner
Copy Link

@geoHeil commented on GitHub (Jul 15, 2025):

Is there any update on this by now?

<!-- gh-comment-id:3073395524 --> @geoHeil commented on GitHub (Jul 15, 2025): Is there any update on this by now?
GiteaMirror commented 2026-04-28 02:17:01 -05:00
Author
Owner
Copy Link

@dglidden commented on GitHub (Mar 11, 2026):

Adding another voice to having some kind of mechanism (env var, command line param, something) that would allow ollama to "redirect" pulling from a local Artifactory of some sort. We use JFrog where I work and the ollama registry is basically blocked, leaving only going through the local JFrog instance to be able to pull models. (However they absolutely support pulling models through it, just which ones based on corporate rules.)

<!-- gh-comment-id:4042262176 --> @dglidden commented on GitHub (Mar 11, 2026): Adding another voice to having some kind of mechanism (env var, command line param, something) that would allow ollama to "redirect" pulling from a local Artifactory of some sort. We use JFrog where I work and the ollama registry is basically blocked, leaving only going through the local JFrog instance to be able to pull models. (However they absolutely support pulling models through it, just which ones based on corporate rules.)
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label question
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#46957
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?