How to Use Local / Remote Ollama with PowerShell 7 on Windows

How to Use Local / Remote Ollama with PowerShell 7 on Windows

This guide shows you how to set up a custom PowerShell command to interact with your local/remote Ollama server from the Windows console.

---

Step 1: Install PowerShell 7

• Go to PowerShell GitHub Releases.
  
• Download the latest MSI installer for Windows (e.g., PowerShell-7.x.x-win-x64.msi).
  
• Run the installer and follow the prompts.
  
• Start PowerShell 7 by typing pwsh in your Start menu or terminal.
  

Step 2: Set Up Your PowerShell Profile

• Open PowerShell 7 (pwsh).
  
• Type:
  
  Code:

  notepad $PROFILE
• If prompted, create the file.
  
• Paste the following function into the file: (replace model and uri with custom parameters)
  
  Code:

  function ollama {
    param(
        [Parameter(ValueFromRemainingArguments=$true)]
        $PromptArgs
    )
    $prompt = $PromptArgs -join " "
    $body = @{ model = "llama3.2"; prompt = $prompt } | ConvertTo-Json
    $response = Invoke-RestMethod -Uri "http://192.168.1.69:7869/api/generate" -Method POST -Body $body -ContentType "application/json"
    $lines = $response -split "`n"
    $answer = ($lines | ForEach-Object {
        try { ($_.Trim() | ConvertFrom-Json).response } catch {}
    }) -join ""
    Write-Host $answer
}
• Save and close Notepad.
  
• Reload your profile by typing:
  
  Code:

  . $PROFILE

Note*
When you define a function (like ollama) in your profile, you can call it by typing its name in the console—this is known as a function-based command in PowerShell.

Step 3: Use Your Custom Command (Or alias)

• In PowerShell 7, type:
  
  Code:

  ollama Your prompt here
• Example:
  
  Code:

  ollama What is the capital of France?
• You’ll see the combined response from your Ollama server.
  

Notes:

• Make sure your Ollama server is running and accessible at the IP and port you specified.
  
• You can change the model name or server address in the function as needed.
  
• This command only sends text prompts and returns text responses.