docs: add preview images to each module readme

Author: ppiova

01-quickstart/README.md

@@ -2,6 +2,8 @@
 
 Pull a model, run a one-shot prompt and discover the OpenAI-compatible API endpoint, all in a single script.
 
+![Quickstart script output](../assets/01-quickstart.svg)
+
 ## Prerequisites
 
 - Docker Desktop 4.40+ with Docker Model Runner enabled

02-openai-api/README.md

@@ -7,6 +7,8 @@ Docker Model Runner exposes an OpenAI-compatible API, so any OpenAI client or pl
 
 No API key is required, but OpenAI client libraries usually expect a non-empty one, so pass any placeholder value.
 
+![curl examples output, non-streaming and streaming](../assets/02-openai-api.svg)
+
 ## curl
 
 ```bash

03-dotnet-chat/README.md

@@ -5,6 +5,8 @@ Docker Model Runner, using the official [`OpenAI`](https://www.nuget.org/package
 
 Because DMR is OpenAI-compatible, the only DMR-specific detail is the base URL.
 
+![.NET streaming chat session](../assets/03-dotnet-chat.svg)
+
 ## Prerequisites
 
 - .NET 10 SDK

04-compose/README.md

@@ -2,6 +2,8 @@
 
 Provision the model and a small app together with a single `docker compose up`.
 
+![docker compose up and a POST /chat response](../assets/04-compose.svg)
+
 The [`compose.yaml`](compose.yaml) uses the top-level `models` element so Compose makes
 `ai/gemma3` available through Docker Model Runner, and binds it to the `chat-api` service
 using the long syntax. Compose injects two environment variables into the container:

05-blazor-chat/README.md

@@ -6,7 +6,7 @@ The OpenAI SDK runs server-side (the same pattern as [03-dotnet-chat](../03-dotn
 [04-compose](../04-compose)), and the model streams its answer token by token to the browser over
 the Blazor SignalR connection. No CORS setup, no API key, no data leaving your machine.
 
-<!-- ![Blazor chat screenshot](../assets/blazor-chat.png) -->
+![Blazor chat UI](../assets/05-blazor-chat.svg)
 
 ## Features