github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #1006] Mobile support #26250

New Issue
Open
opened 2026-04-22 02:21:36 -05:00 by GiteaMirror · 21 comments
GiteaMirror commented 2026-04-22 02:21:36 -05:00
Owner
Copy Link

Originally created by @mikestaub on GitHub (Nov 5, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/1006

Is there a plan to deploy this to iOS or Android so users can run models locally on their mobile devices?

What would it take to achieve this?

Originally created by @mikestaub on GitHub (Nov 5, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/1006 Is there a plan to deploy this to iOS or Android so users can run models locally on their mobile devices? What would it take to achieve this?
GiteaMirror added the feature request label 2026-04-22 02:21:36 -05:00
GiteaMirror commented 2026-04-22 02:21:37 -05:00
Author
Owner
Copy Link

@pepperoni21 commented on GitHub (Nov 5, 2023):

I don't think models would run well on mobile devices, but why not consider developing a mobile app that can connect to a remote Ollama server? But that would be up to the community

<!-- gh-comment-id:1793775731 --> @pepperoni21 commented on GitHub (Nov 5, 2023): I don't think models would run well on mobile devices, but why not consider developing a mobile app that can connect to a remote Ollama server? But that would be up to the community
GiteaMirror commented 2026-04-22 02:21:38 -05:00
Author
Owner
Copy Link

@igorschlum commented on GitHub (Nov 5, 2023):

I would prefer to have a Mac with all LLMs on it and that this Mac is privately accessible over the network through my Internet Service Provider. I would also like to be able to send requests to my desktop Mac from my mobile phone with an HTTP request.

<!-- gh-comment-id:1793785279 --> @igorschlum commented on GitHub (Nov 5, 2023): I would prefer to have a Mac with all LLMs on it and that this Mac is privately accessible over the network through my Internet Service Provider. I would also like to be able to send requests to my desktop Mac from my mobile phone with an HTTP request.
GiteaMirror commented 2026-04-22 02:21:38 -05:00
Author
Owner
Copy Link

@manish7017 commented on GitHub (Nov 6, 2023):

Hey @mikestaub and @pepperoni21, I tried to run the model on my mobile and it ran smoothly.

Btw Device is Realme X2 8gb varient.

Screenshot_2023-11-06-22-05-49-09_84d3000e3f4017145260f7618db1d683
Screenshot_2023-11-06-22-05-25-27_84d3000e3f4017145260f7618db1d683
Screenshot_2023-11-06-22-06-02-52_84d3000e3f4017145260f7618db1d683

<!-- gh-comment-id:1795600217 --> @manish7017 commented on GitHub (Nov 6, 2023): Hey @mikestaub and @pepperoni21, I tried to run the model on my mobile and it ran smoothly. Btw Device is Realme X2 8gb varient. ![Screenshot_2023-11-06-22-05-49-09_84d3000e3f4017145260f7618db1d683](https://github.com/jmorganca/ollama/assets/51908018/67c81a61-168a-4746-b8e6-1de39f5b5466) ![Screenshot_2023-11-06-22-05-25-27_84d3000e3f4017145260f7618db1d683](https://github.com/jmorganca/ollama/assets/51908018/7cc41fbf-362e-41d4-b80e-7397c318823d) ![Screenshot_2023-11-06-22-06-02-52_84d3000e3f4017145260f7618db1d683](https://github.com/jmorganca/ollama/assets/51908018/002c14a8-0716-4357-9b48-3b67118173c4)
GiteaMirror commented 2026-04-22 02:21:39 -05:00
Author
Owner
Copy Link

@pepperoni21 commented on GitHub (Nov 6, 2023):

Hey @mikestaub and @pepperoni21, I tried to run the model on my mobile and it ran smoothly.

Btw Device is Realme X2 8gb varient.

Screenshot_2023-11-06-22-05-49-09_84d3000e3f4017145260f7618db1d683 Screenshot_2023-11-06-22-05-25-27_84d3000e3f4017145260f7618db1d683 Screenshot_2023-11-06-22-06-02-52_84d3000e3f4017145260f7618db1d683

Great, then why not, but again I think that should be made as a community integration like the other frontend apps that have been made for ollama.

<!-- gh-comment-id:1795743458 --> @pepperoni21 commented on GitHub (Nov 6, 2023): > Hey @mikestaub and @pepperoni21, I tried to run the model on my mobile and it ran smoothly. > > Btw Device is Realme X2 8gb varient. > > ![Screenshot_2023-11-06-22-05-49-09_84d3000e3f4017145260f7618db1d683](https://user-images.githubusercontent.com/51908018/280796836-67c81a61-168a-4746-b8e6-1de39f5b5466.jpg) ![Screenshot_2023-11-06-22-05-25-27_84d3000e3f4017145260f7618db1d683](https://user-images.githubusercontent.com/51908018/280796827-7cc41fbf-362e-41d4-b80e-7397c318823d.jpg) ![Screenshot_2023-11-06-22-06-02-52_84d3000e3f4017145260f7618db1d683](https://user-images.githubusercontent.com/51908018/280796842-002c14a8-0716-4357-9b48-3b67118173c4.jpg) Great, then why not, but again I think that should be made as a community integration like the other frontend apps that have been made for ollama.
GiteaMirror commented 2026-04-22 02:21:39 -05:00
Author
Owner
Copy Link

@breitburg commented on GitHub (Nov 11, 2023):

MLC LLM already allows you to run models locally, and it has an API to embed it into your own apps. Furthermore, they already have an app that can download any models from Hugging Face and run them locally.

<!-- gh-comment-id:1806646809 --> @breitburg commented on GitHub (Nov 11, 2023): [MLC LLM](https://llm.mlc.ai/) already allows you to run models locally, and it has an API to embed it into your own apps. Furthermore, they already have an app that can download any models from [Hugging Face](https://huggingface.co/) and run them locally.
GiteaMirror commented 2026-04-22 02:21:40 -05:00
Author
Owner
Copy Link

@danemadsen commented on GitHub (Nov 29, 2023):

For anyone still interested Maid is a mobile app and has support for connecting to an Ollama server remotely:
https://github.com/MaidFoundation/maid

Additionally though, I'm looking into the possibility of using Ollama as a backend to run models locally as well. I'm currently using a hacked together implementation of Llama.cpp which isn't very performant and is constantly breaking. If there's a way to integrate Ollama into my app that would be ideal. From my research Go code can be compiled to run along side flutter so i think its possible, just not very easy.

<!-- gh-comment-id:1832824214 --> @danemadsen commented on GitHub (Nov 29, 2023): For anyone still interested Maid is a mobile app and has support for connecting to an Ollama server remotely: https://github.com/MaidFoundation/maid Additionally though, I'm looking into the possibility of using Ollama as a backend to run models locally as well. I'm currently using a hacked together implementation of Llama.cpp which isn't very performant and is constantly breaking. If there's a way to integrate Ollama into my app that would be ideal. From my research Go code can be compiled to run along side flutter so i think its possible, just not very easy.
GiteaMirror commented 2026-04-22 02:21:41 -05:00
Author
Owner
Copy Link

@vikesh001 commented on GitHub (Feb 6, 2024):

Hey @mikestaub and @pepperoni21, I tried to run the model on my mobile and it ran smoothly.

Btw Device is Realme X2 8gb varient.

Screenshot_2023-11-06-22-05-49-09_84d3000e3f4017145260f7618db1d683 Screenshot_2023-11-06-22-05-25-27_84d3000e3f4017145260f7618db1d683 Screenshot_2023-11-06-22-06-02-52_84d3000e3f4017145260f7618db1d683

How to run ollama in termux

<!-- gh-comment-id:1930541661 --> @vikesh001 commented on GitHub (Feb 6, 2024): > Hey @mikestaub and @pepperoni21, I tried to run the model on my mobile and it ran smoothly. > > Btw Device is Realme X2 8gb varient. > > ![Screenshot_2023-11-06-22-05-49-09_84d3000e3f4017145260f7618db1d683](https://private-user-images.githubusercontent.com/51908018/280796836-67c81a61-168a-4746-b8e6-1de39f5b5466.jpg?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.pc0JKjSTOWox27yJCoC9MHM_9-9bKurle_MWNXpVUfw) ![Screenshot_2023-11-06-22-05-25-27_84d3000e3f4017145260f7618db1d683](https://private-user-images.githubusercontent.com/51908018/280796827-7cc41fbf-362e-41d4-b80e-7397c318823d.jpg?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.ynmczMe1ZebRYoJt0l4Vfou4q_OthxDP191khVuFkAw) ![Screenshot_2023-11-06-22-06-02-52_84d3000e3f4017145260f7618db1d683](https://private-user-images.githubusercontent.com/51908018/280796842-002c14a8-0716-4357-9b48-3b67118173c4.jpg?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.2nKoRD5-4ehZM8pv61-Dk4lbJMEfXtmvpI64QRyITLc) How to run ollama in termux
GiteaMirror commented 2026-04-22 02:21:42 -05:00
Author
Owner
Copy Link

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

How to run ollama in termux

don't you think that if people can install Kalin in termux, then surely they can install or compile ollama?

<!-- gh-comment-id:1949072866 --> @mahavatara commented on GitHub (Feb 16, 2024): > How to run ollama in termux don't you think that if people can install Kalin in termux, then surely they can install or compile ollama?
GiteaMirror commented 2026-04-22 02:21:43 -05:00
Author
Owner
Copy Link

@seanmavley commented on GitHub (Feb 21, 2024):

@vikesh001

How to run ollama in termux

Same as you'd run on your Linux. Copy paste the install script command from https://ollama.com/download

curl -fsSL https://ollama.com/install.sh | sh

Popular terminal app on Android to use is Termux

<!-- gh-comment-id:1958396950 --> @seanmavley commented on GitHub (Feb 21, 2024): @vikesh001 > How to run ollama in termux Same as you'd run on your Linux. Copy paste the install script command from https://ollama.com/download `curl -fsSL https://ollama.com/install.sh | sh` Popular terminal app on Android to use is Termux
GiteaMirror commented 2026-04-22 02:21:43 -05:00
Author
Owner
Copy Link

@bibutikoley commented on GitHub (Feb 28, 2024):

Are there any plans to create mobile libs with kotlin and swift support?

<!-- gh-comment-id:1969424495 --> @bibutikoley commented on GitHub (Feb 28, 2024): Are there any plans to create mobile libs with kotlin and swift support?
GiteaMirror commented 2026-04-22 02:21:44 -05:00
Author
Owner
Copy Link

@suoko commented on GitHub (Mar 3, 2024):

@vikesh001

How to run ollama in termux

Same as you'd run on your Linux. Copy paste the install script command from https://ollama.com/download

curl -fsSL https://ollama.com/install.sh | sh

Popular terminal app on Android to use is Termux

But it requires sudo. Can it be done without root access?

<!-- gh-comment-id:1975244575 --> @suoko commented on GitHub (Mar 3, 2024): > @vikesh001 > > > How to run ollama in termux > > Same as you'd run on your Linux. Copy paste the install script command from https://ollama.com/download > > `curl -fsSL https://ollama.com/install.sh | sh` > > Popular terminal app on Android to use is Termux But it requires sudo. Can it be done without root access?
GiteaMirror commented 2026-04-22 02:21:45 -05:00
Author
Owner
Copy Link

@dbedoyat commented on GitHub (Jul 26, 2024):

https://gitlab.com/-/snippets/3682973

With this tutorial you can run termux on android.

Works well on S23 Ultra
Screenshot_20240725_213811_Termux

<!-- gh-comment-id:2251855843 --> @dbedoyat commented on GitHub (Jul 26, 2024): https://gitlab.com/-/snippets/3682973 With this tutorial you can run termux on android. Works well on S23 Ultra ![Screenshot_20240725_213811_Termux](https://github.com/user-attachments/assets/59057590-ae33-4ef6-b589-66d9dad8efed)
GiteaMirror commented 2026-04-22 02:21:46 -05:00
Author
Owner
Copy Link

@cbellosoto commented on GitHub (Aug 11, 2024):

https://github.com/JHubi1/ollama-app is another very good option that seems better than maid for a more generic use.

<!-- gh-comment-id:2282504983 --> @cbellosoto commented on GitHub (Aug 11, 2024): https://github.com/JHubi1/ollama-app is another very good option that seems better than maid for a more generic use.
GiteaMirror commented 2026-04-22 02:21:46 -05:00
Author
Owner
Copy Link

@knyipab commented on GitHub (Sep 15, 2024):

I packaged ollama to TUR (in openblas cpu backend). It is more readily available and will update along with your pkg update in the future. First make sure you have tur-repo installed by running pkg install tur-repo. Then run:

pkg update
pkg install -y ollama

My 8gen2 runs qwen2:1.5b with 7 tokens/s. You may test yours with ollama serve and then this command:

ollama run --verbose qwen2:1.5b
<!-- gh-comment-id:2351508345 --> @knyipab commented on GitHub (Sep 15, 2024): I packaged `ollama` to [TUR (in openblas cpu backend)](https://github.com/termux-user-repository/tur/pull/1198). It is more readily available and will update along with your `pkg update` in the future. First make sure you have `tur-repo` installed by running `pkg install tur-repo`. Then run: ``` pkg update pkg install -y ollama ``` My 8gen2 runs qwen2:1.5b with 7 tokens/s. You may test yours with `ollama serve` and then this command: ``` ollama run --verbose qwen2:1.5b ```
GiteaMirror commented 2026-04-22 02:21:47 -05:00
Author
Owner
Copy Link

@gabe-l-hart commented on GitHub (Oct 22, 2024):

Has anyone run into the noexec error that @dhiltgen added in #4241 when trying to run on Android inside Termux? I'm able to manually run the runner from the Termux command line (setting LD_LIBRARY_PATH to point to the extracted runner's directory), but when ollama attempts to run the subprocess, it seems to be hitting this clause and erroring out.

<!-- gh-comment-id:2430336845 --> @gabe-l-hart commented on GitHub (Oct 22, 2024): Has anyone run into the `noexec` error that @dhiltgen added in #4241 when trying to run on Android inside Termux? I'm able to manually run the `runner` from the Termux command line (setting `LD_LIBRARY_PATH` to point to the extracted runner's directory), but when `ollama` attempts to run the subprocess, it seems to be hitting [this clause](https://github.com/ollama/ollama/blob/main/llm/server.go#L409) and erroring out.
GiteaMirror commented 2026-04-22 02:21:49 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Oct 22, 2024):

@gabe-l-hart our current linux strategy is to carry the runners as payloads inside the main binary. If that's problematic in a mobile scenario, you can take a look at how we're building the container image in the Dockerfile for inspiration. As long as the ./build/ tree does not have runners at the time you go build and you lay out the runners in ../lib/ollama/runners/... relative to the location of the ollama executable, then you can bypass the payload logic.

<!-- gh-comment-id:2430377447 --> @dhiltgen commented on GitHub (Oct 22, 2024): @gabe-l-hart our current linux strategy is to carry the runners as payloads inside the main binary. If that's problematic in a mobile scenario, you can take a look at how we're building the container image in the Dockerfile for inspiration. As long as the `./build/` tree does not have runners at the time you `go build` and you lay out the runners in `../lib/ollama/runners/...` relative to the location of the ollama executable, then you can bypass the payload logic.
GiteaMirror commented 2026-04-22 02:21:50 -05:00
Author
Owner
Copy Link

@gabe-l-hart commented on GitHub (Oct 22, 2024):

Ah, I'll dig in there. I was just looking over the extraction logic. Thanks!

<!-- gh-comment-id:2430380181 --> @gabe-l-hart commented on GitHub (Oct 22, 2024): Ah, I'll dig in there. I was just looking over the extraction logic. Thanks!
GiteaMirror commented 2026-04-22 02:21:51 -05:00
Author
Owner
Copy Link

@aibangjuxin commented on GitHub (Oct 25, 2024):

I think this requirement is rather necessary, especially for iPad users, because iPads now have 16GB of memory.

<!-- gh-comment-id:2436824715 --> @aibangjuxin commented on GitHub (Oct 25, 2024): I think this requirement is rather necessary, especially for iPad users, because iPads now have 16GB of memory.
GiteaMirror commented 2026-04-22 02:21:51 -05:00
Author
Owner
Copy Link

@Rudxain commented on GitHub (Feb 3, 2025):

https://gitlab.com/-/snippets/3682973

Why does the tutorial use go generate instead of cmake?

And why do the official docs imply that cmake isn't necessary, but then proceed to be like "actually, you need cmake!". Those docs need some serious improvement, I find them misleading and frustrating

I just realized there's a distinction: run is only used for serve; while cmake, go generate, and go build, are used for the CLI app. The docs should be more explicit about that. Now I've noticed they're the same executable file

<!-- gh-comment-id:2629692199 --> @Rudxain commented on GitHub (Feb 3, 2025): > https://gitlab.com/-/snippets/3682973 Why does the tutorial use `go generate` [instead of `cmake`](https://github.com/ollama/ollama/blob/ad22ace439eb3fab7230134e56bb6276a78347e4/docs/development.md#linux)? And why do the official docs imply that `cmake` isn't necessary, but then proceed to be like "actually, you need cmake!". Those docs **need some serious improvement**, I find them misleading and frustrating ~I just realized there's a distinction: `run` is only used for `serve`; while `cmake`, `go generate`, and `go build`, are used for the CLI app.~ The docs should be more explicit about that. Now I've noticed they're the same executable file
GiteaMirror commented 2026-04-22 02:21:52 -05:00
Author
Owner
Copy Link

@mcandre commented on GitHub (May 10, 2025):

Mobile support unlocks so many useful features...

Decouple workstations, support tablet and phone operation

Inform and entertain where Internet access is limited

Compete with other mobile AI systems

Empower privacy efforts on mobile

Reduce antenna power consumption (though GPU consumption increases)

Optionally decouple from remote ollama servers

Promote small models

<!-- gh-comment-id:2868401759 --> @mcandre commented on GitHub (May 10, 2025): Mobile support unlocks so many useful features... Decouple workstations, support tablet and phone operation Inform and entertain where Internet access is limited Compete with other mobile AI systems Empower privacy efforts on mobile Reduce antenna power consumption (though GPU consumption increases) Optionally decouple from remote ollama servers Promote small models
GiteaMirror commented 2026-04-22 02:21:52 -05:00
Author
Owner
Copy Link

@Eloitor commented on GitHub (Dec 13, 2025):

https://github.com/google-ai-edge/gallery
With this app you can use gemma3, gemma-3n, qwen-2.5, phi-4-mini and DeepSeek-R1-Distill-Qwen

<!-- gh-comment-id:3649151930 --> @Eloitor commented on GitHub (Dec 13, 2025): https://github.com/google-ai-edge/gallery With this app you can use gemma3, gemma-3n, qwen-2.5, phi-4-mini and DeepSeek-R1-Distill-Qwen
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 feature request
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#26250
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?