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 #2388] Local caching / Custom Registry #27149

New Issue
Closed
opened 2026-04-22 04:07:37 -05:00 by GiteaMirror · 8 comments
GiteaMirror commented 2026-04-22 04:07:37 -05:00
Owner
Copy Link

Originally created by @KaruroChori on GitHub (Feb 7, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2388

Is there any supported solution to:

  • locally host a custom registry?
  • at the very least a partial mirror of the registry for those models which have been downloaded once already?

Downloading the same model from the original source on multiple machines within the same network is quite slow and wasteful.

Originally created by @KaruroChori on GitHub (Feb 7, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2388 Is there any supported solution to: - locally host a custom registry? - at the very least a partial mirror of the registry for those models which have been downloaded once already? Downloading the same model from the original source on multiple machines within the same network is quite slow and wasteful.
GiteaMirror commented 2026-04-22 04:07:40 -05:00
Author
Owner
Copy Link

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

Great question. I have similar issue.

<!-- gh-comment-id:1947606792 --> @arkohut commented on GitHub (Feb 16, 2024): Great question. I have similar issue.
GiteaMirror commented 2026-04-22 04:07:41 -05:00
Author
Owner
Copy Link

@easp commented on GitHub (Feb 23, 2024):

https://github.com/ollama/ollama/issues/914#issuecomment-1953482174

<!-- gh-comment-id:1960721357 --> @easp commented on GitHub (Feb 23, 2024): https://github.com/ollama/ollama/issues/914#issuecomment-1953482174
GiteaMirror commented 2026-04-22 04:07:41 -05:00
Author
Owner
Copy Link

@KaruroChori commented on GitHub (Feb 23, 2024):

Thanks for reporting.
So it appears like, under the hood, it works like a docker repository, not just as flavour.

<!-- gh-comment-id:1960946158 --> @KaruroChori commented on GitHub (Feb 23, 2024): Thanks for reporting. So it appears like, under the hood, it works like a docker repository, not just as flavour.
GiteaMirror commented 2026-04-22 04:07:42 -05:00
Author
Owner
Copy Link

@bmizerany commented on GitHub (Mar 11, 2024):

Ollama keeps a cache of your already download models under ~/.ollama or $OLLAMA_MODELS if set.

Ollama is currently compatible with Docker Registry, but we do not guarantee it will remain compatible.

It should be possible to host your own registry with authentication turned off (we implemented custom auth for our hosted version) for use as a private registry.

Closing - please reopen if you have any further questions.

<!-- gh-comment-id:1989307410 --> @bmizerany commented on GitHub (Mar 11, 2024): Ollama keeps a cache of your already download models under `~/.ollama` or `$OLLAMA_MODELS` if set. Ollama is currently compatible with Docker Registry, but we do not guarantee it will remain compatible. It should be possible to host your own registry with authentication turned off (we implemented custom auth for our hosted version) for use as a private registry. Closing - please reopen if you have any further questions.
GiteaMirror commented 2026-04-22 04:07:43 -05:00
Author
Owner
Copy Link

@samheutmaker commented on GitHub (May 9, 2024):

@bmizerany Do you have any documentation that explains Docker registry compatibility? We're trying to set up a LAN registry for our team to share models. We will open source anything we build around this for re-use by the community.

<!-- gh-comment-id:2103477896 --> @samheutmaker commented on GitHub (May 9, 2024): @bmizerany Do you have any documentation that explains Docker registry compatibility? We're trying to set up a LAN registry for our team to share models. We will open source anything we build around this for re-use by the community.
GiteaMirror commented 2026-04-22 04:07:44 -05:00
Author
Owner
Copy Link

@bmizerany commented on GitHub (May 13, 2024):

@samheutmaker Currently, we do not have an official registry API, which is why we don't have any documentation available for Docker registry compatibility.

We are considering and working towards a solution for remote caching. There are no dates for its release currently.

For now, we unofficially use the GET layer, POST/PUT(upload) layer and read/commit manifests.

<!-- gh-comment-id:2107657364 --> @bmizerany commented on GitHub (May 13, 2024): @samheutmaker Currently, we do not have an official registry API, which is why we don't have any documentation available for Docker registry compatibility. We are considering and working towards a solution for remote caching. There are no dates for its release currently. For now, we unofficially use the GET layer, POST/PUT(upload) layer and read/commit manifests.
GiteaMirror commented 2026-04-22 04:07:45 -05:00
Author
Owner
Copy Link

@haydonryan commented on GitHub (Oct 20, 2024):

Plus one on some local registry. I run both pacoloco and Panamax to mirror stuff locally (and nightly at 3am). Would love to do cache models too as I have a team who use the same models on their machines.

<!-- gh-comment-id:2425189346 --> @haydonryan commented on GitHub (Oct 20, 2024): Plus one on some local registry. I run both pacoloco and Panamax to mirror stuff locally (and nightly at 3am). Would love to do cache models too as I have a team who use the same models on their machines.
GiteaMirror commented 2026-04-22 04:07:45 -05:00
Author
Owner
Copy Link

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

@samheutmaker Currently, we do not have an official registry API, which is why we don't have any documentation available for Docker registry compatibility.

We are considering and working towards a solution for remote caching. There are no dates for its release currently.

For now, we unofficially use the GET layer, POST/PUT(upload) layer and read/commit manifests.

@samheutmaker Is there now a plan for when the API will be Docker API compatible so that the registry can be cached locally via Sonatype Nexus or JFrog Artefactory via proxy repository?

Same issue as #5298 ?

<!-- gh-comment-id:2579361655 --> @cniweb commented on GitHub (Jan 9, 2025): > @samheutmaker Currently, we do not have an official registry API, which is why we don't have any documentation available for Docker registry compatibility. > > We are considering and working towards a solution for remote caching. There are no dates for its release currently. > > For now, we unofficially use the GET layer, POST/PUT(upload) layer and read/commit manifests. @samheutmaker Is there now a plan for when the API will be Docker API compatible so that the registry can be cached locally via Sonatype Nexus or JFrog Artefactory via proxy repository? Same issue as #5298 ?
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
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#27149
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?