Home Creators Posts Import Register Favorites Logout

Downloads

Content

Hey everyone! I've created a 1-click installer for the AMAZING SillyTavern UI for your Phone and Mobile devices. Keep in mind that this was made to work with Termux on Android so IOS users might have to try the manual install since I personally don't have access to any Apple products to try it out myself...sorry about that. In any case, you Absolutely need to try this out!

Watch the video on SillyTavern for Phones: https://youtu.be/kTAKL97FL8g

The installer of course automates the entire install process!

Here's how it works:

1. Download the installer (silly.sh) and the launcher (launchsilly.sh). 2. Create a new folder on your internal storage phone called "AI" and place those 2 files inside 3. Download, install and run Termux 4. Run the following commands:termux-setup-storage cd ~/storage/shared/AI bash silly.sh 5. If it asks for an action just press  "y" and then press enter (might ask you multiple times) 6. Wait for everything to be finished 7. Choose the browser to run sillytavern with 8. Use my Google colab doc: https://colab.research.google.com/drive/1DeTTwJPjmZMJNYCzYHX6LfsUay9-s0Kk?usp=sharing

then use the public "OpenAI-compatible Api key" inside sillytavern

9.???

10. Profit  😎

You can also use the extras-APi Google Colab doc to have access to a bunch of other api options and extensions: https://colab.research.google.com/github/SillyTavern/SillyTavern/blob/release/colab/GPU.ipynb?authuser=2

What's great is that you can use those Google colab on your PC or any mobile devices and have pretty incredible conversation speed, so that even if you have a weird gpu or no GPU at all you can enjoy sexy-shmexy roleplay whenever and wherever you are 😉

As always, supporting me on Patreon allows me to keep creating helpful resources like this for the community. Thank you for your support 😉!

Files

Comments

hammo 740

If anyone has tried to manual install this using iPhone pls let me know

Xavier Shipley

both launchsilly.sh and silly.sh sites can't be reached error?

Joel Rivera

Couldn't get this to work on Google Pixel 8 Pro android phone, doesn't seem to play nice with the linux terminal app

Zolon's Shop of WonderZ

termux-setup-storage - /data/data/com.termux/files/usr/bin/termux-setup-storage : line 24: 17694 Aborted

Zolon's Shop of WonderZ

If you are on Android 14, you also need to go into the app permissions for Termux and give it full file access. After I did that, the pkg install line worked, which then allowed the storage command to run.

Aitrepreneur

These files should be downloaded without any issues, just tried it

Ma No

Interesting Youtube comment advised setting the AI up on your computer and Remoting in to the desktop. Might be something worth of exploring for a video.

Aitrepreneur

you can do that but it's not as interesting since first, you need a computer always running while this allows you to be anywhere you want and as long as you have internet connection, it will work as is

Ma No

That's fair just thought it might make an interesting video. Or maybe not just RDP into the desktop. Wasn't sure if there was a good solution. Do you know if the characters/roleplays you have can be carried over/ported to your mobile device or would you have to start over or export/import your current roleplay?

Nabil Boulezaz

Can we use stable diffusion on Android?

Aitrepreneur

not really, except either using a Google colab doc or running SD on a pc then connecting to the url with your phone.

BP

I just tried downloading on my phone via Firefox browser and it said download failed. Tapped and long pressed, both failed

Anonymous

Both chrome and edge on a Samsung fold say "This site can’t be reachedCheck if there is a typo in launchsilly.sh. DNS_PROBE_FINISHED_NXDOMAIN" and fail.

AdAstra84

I figured it already out with the permission in Android 14 but still getting failures when trying following the steps. Got any ideas? It says the same for installing termux-am as for the silly tavern package... verification errors and because of it no installation is possible. I granted termux any rights and allowed unsafe installations. Nothing helped.

AdAstra84

Errors are: failed to fetch and the repository is not signed

BP

Both silly.sh and launchsilly.sh link to their filenames not their URLs.

Nabil Boulezaz

Can i transfer my progress when i finally get gpu ?

Aitrepreneur

you can, all your chats should be saved in the backups folder

baku

same issues. do not load, just errors on both

baku

fixed it by downloading the files on my pc, sending it to my phone

Peter Bernaiche

I'm getting the following error when trying g to enter the first command after installing termux.: Report issues at https://termux.com/issues ~ $ termux-setup-storage /data/data/com.termux/files/usr/bin/termux-setup-storage: line 29: 29934 Aborted am broadcast --user 0 --es com.termux.app.reload_style storage -a com.termux.app.reload_style com.termux > /dev/null

SilentDenizen

is it at all possible to use a low accuracy .gguf on text generation web ui locally on android?

Aitrepreneur

make sure that you have enabled all permissions for termux in the apps settings, on some phone for some reason you need to do that

SilentDenizen

the Samsung Galaxy s23 ultra able to? I would love to have it local. and if so, how would it be done? Winlator or some other windows emulator?

Aitrepreneur

Might be possible to install the textgen webui on the phone directly using termux also, I haven't tried it yet myself but tbh the text gen would be so slow I would not advise doing that, even modern pc struggles to generate at a good speed so imagine a phone..

SilentDenizen

ah. I hope in the next few years that technology and programming advance far enough for it to be possible. I dislike how the doc provided has a limit on it. I blew through it before the day was gone just exploring different AI character models. I instead decided to have my PC at home run just the textgen web ui and opted to share and used the link generated to use textgen from my phone. I do not know how to get the gradiolive link to work with sillytavern.

Jay

Facing the same issue over and over again ~ $ termux-setup-storage /data/data/com.termux/files/usr/bin/termux-setup-storage: line 29: 29934 Aborted am broadcast --user 0 --es com.termux.app.reload_style storage -a com.termux.app.reload_style com.termux > /dev/null Even after giving all permissions for termux in the apps settings

Aitrepreneur

ok so first, if you are sure that you gave termux all persmissions, does everything work when launching the first few command? termux-setup-storage cd ~/storage/shared/AI After running the first command, termux should have asked you persmissions to access your phone folder and files. And the second command should have taken you to the Ai folder on your Phone where you placed the 2 files (installer and launcher)

Steven Sturgill

Samsung Galaxy Note Fold4 user All permissions given App restarted Silly.sh and Silly installer both in /Internal storage/AI Command I wrote ~ $ termux-setup-storage Message I received /data/data/com.termux/files/usr/bin/termux-setup-storage: line 29: 15731 Aborted am broadcast --user 0 --es com.termux.app.reload_style storage -a com.termux.app.reload_style com.termux > /dev/null Can't even make it to the other commands Attempting: ~ $ CD ~/storage/shared/ai Result: No command CD found, did you mean: Command ab in package apache2 Command at in package at Command bc in package bc Command ar in package binutils Command bk in package bk ....(More irrelevant commands)

Aitrepreneur

ok first run this command: pkg update && pkg upgrade say Y for everything then do the termux-setup-storage command the reason why you got an error with CD is because if the case sensitivity, it's "cd" not "CD" so try this instead: cd ~/storage/shared/AI

Steven Sturgill

Thanks for the reply man, love your videos Attempt: ~ $ pkg update && pkg update Response: Checking availability of current mirror: ok Ign:1 https://deb.kcubeterm.me/termux-main stable InRelease Get:2 https://deb.kcubeterm.me/termux-main stable Release [510 B] Get:3 https://deb.kcubeterm.me/termux-main stable Release.gpg [514 B] Err:3 https://deb.kcubeterm.me/termux-main stable Release.gpg Signed file isn't valid, got 'NODATA' (does the network require authentication?) Reading package lists... Done E: GPG error: https://deb.kcubeterm.me/termux-main stable Release: Signed file isn't valid, got 'NODATA' (does the network require authentication?)

Aitrepreneur

you sure you typed it correctly? it's pkg update && pkg upgrade not pkg update && pkg update you can just do pkg update and then pkg upgrade Next try: apt update apt upgrade

bmxmarine

I get this, Do you want to continue? (y/n) y ~/downloads $ cd ~/storage/emulated/0/Ai bash: cd: /data/data/com.termux/files/home/storage/emulated/0/Ai: No such file or directory ~/downloads $ cd ~/storage/emulated/0/Ai

Aitrepreneur

have you typed all the previous commands correctly? termux-setup-storage press enter (should give you a permission message) cd ~/storage/shared/AI press enter (should move you inside the AI folder you created)

Ryan Clancy

Not sure what's going on... Using the colab doc, for about 2 hours yesterday I was fine, I haven't been able to connect to it since. When I press connect in the plug tab of ST, after copying the api url, it just... Does nothing. Never connects. I'm not even sure what other information I could offer to help you help me. Hahaha

Aitrepreneur

have you selected the T4 gpu on google colab doc? It does not do it automatically, you need to do it manually each time

Ryan Clancy

I forgot initially but went back and changed it. I think that was the issue though, as forcing the colab page and silly tavern closed, and then rerunning the colab and being sure to select t4 before anything else produced a working url. Thanks!

Ryan Clancy

I keep getting "You cannot currently connect to a GPU due to usage limits in Colab. Learn more To get more access to GPUs, consider purchasing Colab compute units with Pay As You Go." when I try to connect to T4. Am I doing something wrong?

Aitrepreneur

well as it says there is a limit to the gpu usage on google colab, just use another google account to go around this restriction

Ryan Clancy

Ahh I see. The restriction is on my end! Gotcha thanks! That's an easy fix. I assumed those were messages for the "host" but I have no idea what's really going on. Hahaha thanks again!

Kubilay göze

is it somehow possible to use my pc as the server and use the sillytavern UI on my phone connected to the same network ?

Aitrepreneur

Yes can just activate the public url in the session tab in the text gen webui and use that url inside sillytavern on your phone

Kubilay göze

I get "Models endpoint is offline." message in termux when I put the public url that ends with '.gradio.live' in sillytavern. when I put the link in my browser, the oobabooga UI loads without issues. What am I missing? I Checked "share" and "public api" in the session tab. still nothing.

Aitrepreneur

That's not the right url, you need to use the "OpenAi-compayible API URL" that ends in trycloudflare.com Check the video again to see what I mean

Nils P.

I just had exactly the same problems. When I try to access the website (which is specified in the code) where the updates are supposed to be downloaded, I am redirected to a ring advertising page for whatever reason. Therefore no update can be pulled from the source. The solution for me was to select a mirror as recommended. For this: termux-change-repo //select any Mirror in the menu Then execute the commands again. Now it should be able to pull the updates and work. apt update apt upgrade pkg update pkg upgrade (Write Y for each request) You should not get any error messages now. After that: termux-setup-storage (No output = no error) cd ~/storage/shared/AI Now everything should work

Nils P.

For those who also have problems running the termux-setup-storage command: I have run the CMD "apt update" as recommended. However, I got an error message that the source is not signed. Solution for me: When I try to access the website (specified in the code) where the updates should be pulled, I am redirected to a, for whatever reason, ring advertising page. Therefore no update can be pulled from the source. The solution is to select a mirror as recommended. For this: termux-change-repo //select any Mirror in the menu Then execute the commands again. Now it should be able to pull the updates and work. apt update apt upgrade pkg update pkg upgrade (Write Y for each request) You should not get any error messages now. After that: termux-setup-storage (No output = no error) cd ~/storage/shared/AI Now everything should work... at least for me now. Hope I could help you.

Aitrepreneur

that's...a weird behaviour O_o but thanks for the info

Steven Sturgill

Okay, so the mirror went through, bunch of stuff installed, all of the apt and pkg update upgrades went through, but when I got to locating the directory I keep getting ~ $ cd ~storage/shared/AI "bash: cd: ~storage/shared/AI: No such file or directory" So I tried using the actual directory path in my phone which is /Internal storage/AI and other guesses and got the same "~ $ cd ~Internalstorage/AI bash: cd: ~Internalstorage/AI: No such file or directory ~ $ cd Internalstorage bash: cd: Internalstorage: No such file or directory ~ $ cd ~Internalstorage/ bash: cd: ~Internalstorage/: No such file or directory"

wanyu jerry

I met same problem, so do i have to change google account every time when i run out of usage?

Jon

Anyone get the "could not extract reply in 5 attempts" error in ST? From what im seeing its a nodejs update issue, but that can't be updated within Termux until they rollout a package update.. thoughts? If it's currently running for anyone else then I'm doing something wrong

Ryan Clancy

Instead of using a Google Colab, can I just set my home desktop's OOBABOOGA up to be accessible anywhere from my phone with an API? Do you have a tutorial on that?

Aitrepreneur

yes, just go to the session tab, and enable openai and public url checkbox, apply the changes and it will then give you a long public url to connect it to sillytavern

Luscious Vampire

Google Colab doc is giving an import error after downloading the llm and it stops there. Also, very early on in the code it says an error loading openai extension which I'm assuming isn't good either since we need it to connect to ST. Just wanted to give a heads up or see if it's a known issue.

Aitrepreneur

make sure that you correctly chose the gpu in google colab doc in "connect", otherwise it's not gonna work

Aitrepreneur

Yup just checked, there was indeed a recent update that made it not work anymore, just updated the colab doc, now it works ;)

SilentDenizen

npm ERR! code ERR_SSL_CIPHER_OPERATION_FAILED npm ERR! 00552A7971000000:error:1C800066:Provider routines:ossl_gcm_stream_update:cipher operation failed:providers/implementations/ciphers/ciphercommon_gcm.c:343: npm ERR!

SilentDenizen

Found a solution. Using Total Commander, tap on the blue + icon, tap on the 3 bars on top, add termux location. Copy silly.sh and launchsilly.sh bypass giving termux storage permissions.

Aitrepreneur

the installers are listed a bit below the post, the first "link" is not a link, it's just Patreon being dumb and thinking silly.sh is a url

Juno

Able to get it working on my phone, it's pretty awesome! As far as chat logs, newly created characters on my phone, imported characters etc... where does it all get saved too? Ive tried to look for like an hour in my phones folders with my PC and on it and can't seem to find it? Is it saved within termux or chrome, or somewhere else?

Aitrepreneur

everything is inside the /data/data/com.termux/files/home/SillyTavern folders

Juno

I must have some settings somewhere locked on my phone then or something that makes it not visible? My data/com.termux/files folder is empty. First place I looked, so that's why I got confused on where it all might be being stored. But I am fairly new in messing around with Android and my phone.

Jon

If you pay for Google colab what's the best gpu for bigger models, T4, A100, or V100? Or should I just stick with the t4 suggestion from the video, also will enabling high ram help? thanks

Eyepatch_101

How do I run the XTTS.bat file when I don't have a GPU? I have colab pro with me, is there a way to create and API link, can you guide??

Jon

Hey Aitrepreneur, can you check the Google doc to see if it's still working? Wondering if they did another update, thanks as always!

Benni

Hey, I run into an error when trying setting this up: npm ERR! code ENOENT npm ERR! syscall open npm ERR! path /data/data/com.termux/files/home/package.json npm ERR! errno -2 npm ERR! enoent ENOENT: no such file or directory, open '/data/data/com.termux/files/home/package.json' npm ERR! enoent This is related to npm not being able to find a file. npm ERR! enoent

P

Could not get to work. Over and over, "could not extract reply in 5 attempts " on silly tavern

AngwySquid

I've been trying to get the open ai url but this keep showing up In collab what should I do? "ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. google-colab 1.0.0 requires pandas==2.0.3, but you have pandas 2.2.2 which is incompatible. imageio 2.31.6 requires pillow<10.1.0,>=8.3.2, but you have pillow 10.3.0 which is incompatible. tensorflow 2.15.0 requires tensorboard<2.16,>=2.15, but you have tensorboard 2.16.2 which is incompatible. Successfully installed DataProperty-1.0.1 GitPython-3.1.43 Pillow-10.3.0 accelerate-0.27.2 aiofiles-23.2.1 aqlm-1.1.3 auto-gptq-0.6.0+cu121 autoawq-0.2.3 autoawq-kernels-0.0.6 bitsandbytes-0.43.1 colorama-0.4.6 coloredlogs-15.0.1 cramjam-2.8.3 datasets-2.19.0 dill-0.3.8 diskcache-5.6.3 docker-pycreds-0.4.0 einops-0.8.0 exllamav2-0.0.19+cu121 fastapi-0.110.3 fastparquet-2024.2.0 ffmpy-0.3.2 flash-attn-2.5.6 gekko-1.1.1 gitdb-4.0.11 gptq-for-llama-0.1.1+cu121 gradio-4.26.0 gradio-client-0.15.1 h11-0.14.0 hqq-0.1.5 httpcore-1.0.5 httpx-0.27.0 huggingface-hub-0.22.2 humanfriendly-10.0 jinja2-3.1.2 jsonlines-4.0.0 llama-cpp-python-0.2.64 llvmlite-0.42.0 lm_eval-0.3.0 mbstrdecoder-1.1.3 multiprocess-0.70.16 ninja-1.11.1.1 numba-0.59.1 numpy-1.26.4 nvidia-cublas-cu12-12.1.3.1 nvidia-cuda-cupti-cu12-12.1.105 nvidia-cuda-nvrtc-cu12-12.1.105 nvidia-cuda-runtime-cu12-12.1.105 nvidia-cudnn-cu12-8.9.2.26 nvidia-cufft-cu12-11.0.2.54 nvidia-curand-cu12-10.3.2.106 nvidia-cusolver-cu12-11.4.5.107 nvidia-cusparse-cu12-12.1.0.106 nvidia-nccl-cu12-2.19.3 nvidia-nvjitlink-cu12-12.4.127 nvidia-nvtx-cu12-12.1.105 openai-1.24.0 optimum-1.17.1 orjson-3.10.1 pandas-2.2.2 pathvalidate-3.2.0 peft-0.8.2 portalocker-2.8.2 psutil-5.9.8 pybind11-2.12.0 pycountry-23.12.11 pydub-0.25.1 pytablewriter-1.2.0 python-multipart-0.0.9 rouge-1.0.1 rouge-score-0.1.2 ruff-0.4.2 sacrebleu-1.5.0 scipy-1.13.0 semantic-version-2.10.0 sentencepiece-0.2.0 sentry-sdk-2.0.1 setproctitle-1.3.3 shellingham-1.5.4 smmap-5.0.1 sqlitedict-2.1.0 tabledata-1.3.3 tcolorpy-0.1.6 tensorboard-2.16.2 timm-0.9.16 tomlkit-0.12.0 tqdm-multiprocess-0.0.11 transformers-4.40.1 typepy-1.3.2 uvicorn-0.29.0 wandb-0.16.6 websockets-11.0.3 xxhash-3.4.1 zstandard-0.22.0 WARNING: The following packages were previously imported in this runtime: [PIL,psutil] You must restart the runtime in order to use newly installed versions."

AngwySquid

I know nothing about coding but I finally got it to connect with the help of gpt 4

philip holman

I am also getting this. after some googling it has something to do with nodeJS version apparently??? but i am not finding a quick solution myself

Senjka Danhieux

In the google colab, I get this: ImportError: /usr/local/lib/python3.10/dist-packages/flash_attn_2_cuda.cpython-310-x86_64-linux-gnu.so: undefined symbol: _ZN3c104cuda9SetDeviceEi

Tridiot

I keep getting the error "Uncaught exception: Error: listen EADDRINUSE: address already in use 127.0.0.1:8000" Then it stops the process

Tridiot

I found the issue, if you have multiple devices on the same network and have the server address the same across the devices then you get the error. Could be wrong though...

James IV

The commands aren't working in Terms. Getting 5108 aborted on the storage setup. And no such time of directory on the second line

Kennedy Vieira Lopes

For me, the following error appears on Google collab. RuntimeError: Failed to import transformers.models.auto.processing_auto because of the following error (look up to see its traceback): module 'torch.library' has no attribute 'register_fake'