Ollama is an innovative Al platform designed to democratize access to large language models (LLMs) by allowing users to run these models locally on their machines. This capability enhances privacy and control over data, as it mitigates concerns associated with cloud-based solutions.

​
Step-by-step installation guide

1

Check if model is available for quick installation

Note that all AI models in Shinkai are powered by Ollama. Before installing Ollama models manually, check the available models list for quick installation.

You can also check in your terminal the installed models in your device:

$ ollama ls
NAME                         ID              SIZE      MODIFIED    
bespoke-minicheck:7b         66607904c165    4.7 GB    4 hours ago    
aya:8b                       7ef8c4942023    4.8 GB    4 hours ago    
llama3.1:8b-instruct-q4_1    e129e608a752    5.1 GB    3 days ago     
snowflake-arctic-embed:xs    bf75350e1752    45 MB     3 days ago
If the model cannot be found in the quick installation list, proceed to step 2.
2

Pull new model from Ollama

Go to Ollama, and browse the model you want to install. Copy the code ollama run <model> and paste it in your terminal.

For example:

$ ollama run bespoke-minicheck
pulling manifest 
pulling 78b0a71988a5... 100% ▕████████████████████████████▏ 4.7 GB                         
pulling 016aea4b585f... 100% ▕████████████████████████████▏  340 B                         
pulling 60ed67c565f8... 100% ▕████████████████████████████▏  506 B                         
pulling 7a3775deddf8... 100% ▕████████████████████████████▏   50 B                         
pulling a870cda2bff0... 100% ▕████████████████████████████▏  495 B                         
verifying sha256 digest 
writing manifest 
success

Check that the model is now installed in your device:

$ ollama ls
NAME                              ID              SIZE      MODIFIED       
bespoke-minicheck:latest          66607904c165    4.7 GB    2 minutes ago     
paraphrase-multilingual:latest    ba13c2e06707    562 MB    11 minutes ago    
bespoke-minicheck:7b              66607904c165    4.7 GB    5 hours ago       
aya:8b                            7ef8c4942023    4.8 GB    5 hours ago       
llama3.1:8b-instruct-q4_1         e129e608a752    5.1 GB    3 days ago        
snowflake-arctic-embed:xs         bf75350e1752    45 MB     3 days ago
3

Add model to Shinkai

Open Shinkai > Go to AIs > + Add AI > + Manually Add AI

    1. Select Ollama in Model Provider.
    1. Select the Model Type you just pulled in step 2.
  • (Optional) You can edit the AI Name —> This is how it will be displayed in your Chats view.
    This field only accepts alphanumeric characters and underscores.
    1. You DO NOT need an API key, as Ollama runs locally on your device.

Congratulations! You have now installed a new Ollama model in your Shinkai Desktop App!

Ollama is an innovative Al platform designed to democratize access to large language models (LLMs) by allowing users to run these models locally on their machines. This capability enhances privacy and control over data, as it mitigates concerns associated with cloud-based solutions.

​
Step-by-step installation guide

1

Check if model is available for quick installation

Note that all AI models in Shinkai are powered by Ollama. Before installing Ollama models manually, check the available models list for quick installation.

You can also check in your terminal the installed models in your device:

$ ollama ls
NAME                         ID              SIZE      MODIFIED    
bespoke-minicheck:7b         66607904c165    4.7 GB    4 hours ago    
aya:8b                       7ef8c4942023    4.8 GB    4 hours ago    
llama3.1:8b-instruct-q4_1    e129e608a752    5.1 GB    3 days ago     
snowflake-arctic-embed:xs    bf75350e1752    45 MB     3 days ago
If the model cannot be found in the quick installation list, proceed to step 2.
2

Pull new model from Ollama

Go to Ollama, and browse the model you want to install. Copy the code ollama run <model> and paste it in your terminal.

For example:

$ ollama run bespoke-minicheck
pulling manifest 
pulling 78b0a71988a5... 100% ▕████████████████████████████▏ 4.7 GB                         
pulling 016aea4b585f... 100% ▕████████████████████████████▏  340 B                         
pulling 60ed67c565f8... 100% ▕████████████████████████████▏  506 B                         
pulling 7a3775deddf8... 100% ▕████████████████████████████▏   50 B                         
pulling a870cda2bff0... 100% ▕████████████████████████████▏  495 B                         
verifying sha256 digest 
writing manifest 
success

Check that the model is now installed in your device:

$ ollama ls
NAME                              ID              SIZE      MODIFIED       
bespoke-minicheck:latest          66607904c165    4.7 GB    2 minutes ago     
paraphrase-multilingual:latest    ba13c2e06707    562 MB    11 minutes ago    
bespoke-minicheck:7b              66607904c165    4.7 GB    5 hours ago       
aya:8b                            7ef8c4942023    4.8 GB    5 hours ago       
llama3.1:8b-instruct-q4_1         e129e608a752    5.1 GB    3 days ago        
snowflake-arctic-embed:xs         bf75350e1752    45 MB     3 days ago
3

Add model to Shinkai

Open Shinkai > Go to AIs > + Add AI > + Manually Add AI

    1. Select Ollama in Model Provider.
    1. Select the Model Type you just pulled in step 2.
  • (Optional) You can edit the AI Name —> This is how it will be displayed in your Chats view.
    This field only accepts alphanumeric characters and underscores.
    1. You DO NOT need an API key, as Ollama runs locally on your device.

Congratulations! You have now installed a new Ollama model in your Shinkai Desktop App!