01.3-hello-chat

Chapter 1.3 -- Hello, IChatClient

Companion code for Generative AI in .NET, Chapter 1 sections 1.3.2 -- 1.3.3.

A console hello-world that calls IChatClient with a small recipe-assistant system prompt, then streams a follow-up turn. Provider is selectable at run time -- Ollama (default, offline), OpenAI direct, or GitHub Models.

Run it

# Default: Ollama with phi4-mini on localhost:11434
dotnet run --project samples/ch01-foundations/01.3-hello-chat

# OpenAI direct
export OPENAI_API_KEY=sk-...
dotnet run --project samples/ch01-foundations/01.3-hello-chat -- openai

# GitHub Models
export GITHUB_TOKEN=ghp_...
dotnet run --project samples/ch01-foundations/01.3-hello-chat -- github

What it shows

• Constructing IChatClient from three different providers behind a single interface.
• Building a List<ChatMessage> with system + user turns.
• Calling both GetResponseAsync (full response) and GetStreamingResponseAsync (token stream).
• Using AddMessages to grow the history correctly between turns.

Manuscript reference

• Manuscript/Chapter-01.md, sections 1.3.2 (project setup) and 1.3.3 (the IChatClient contract).

Prerequisites

• .NET 9 SDK.
• Optional: Ollama running locally with phi4-mini pulled (ollama pull phi4-mini).
• Optional: OPENAI_API_KEY or GITHUB_TOKEN for cloud providers.