Why should you care?

LocalAIDesk’s backend is ASP.NET Core — every feature eventually becomes a service method around this call.

See it live — copy this example

Run CLI steps in a terminal. Run C# samples in a .NET 8 console or Web API project with Ollama running on localhost:11434.

// Program.cs / service
using var client = new HttpClient { BaseAddress = new Uri("http://localhost:11434") };
var payload = new {
  model = "phi3:mini",
  messages = new[] { new { role = "user", content = "Hello from LocalAIDesk" } },
  stream = false
};
var res = await client.PostAsJsonAsync("/api/chat", payload);
res.EnsureSuccessStatusCode();
var json = await res.Content.ReadAsStringAsync();
Console.WriteLine(json);

What happened?

• HttpClient posts to /api/chat.
• PostAsJsonAsync serializes the anonymous object.
• EnsureSuccessStatusCode throws on 4xx/5xx so you notice failures early.

Practice next

1. Create a console or Web API project.
2. Add the HttpClient call.
3. Run with Ollama up.
4. Move model name to appsettings.json.
5. Add a 60s Timeout.

Remember

C# talks to Ollama over HTTP. Configure base URL. Parse the JSON reply carefully.

First LocalAIDesk call

Backend returns a hello string from phi3.

Outcome: End-to-end local path works.