OpenAI Provider

The OpenAI provider integrates Reveal SDK AI with OpenAI's chat completion API, enabling access to models like GPT-4.1, GPT-4o, and reasoning models like o3 and o4.

Installation

Install the OpenAI provider NuGet package:

dotnet add package Reveal.Sdk.AI.OpenAI

Configuration

Basic Setup

Add the OpenAI provider in your Program.cs:

using Reveal.Sdk;
using Reveal.Sdk.AI;

var builder = WebApplication.CreateBuilder(args);

builder.Services.AddControllers().AddReveal();

builder.Services.AddRevealAI()
    .AddOpenAI(options =>
    {
        options.ApiKey = builder.Configuration["RevealAI:OpenAI:ApiKey"];
    });

var app = builder.Build();
app.MapControllers();
app.Run();

Using appsettings.json

The provider automatically binds to the RevealAI:OpenAI configuration section:

appsettings.json

{
  "RevealAI": {
    "OpenAI": {
      "ApiKey": "sk-your-api-key-here",
      "Model": "gpt-4.1",
      "Temperature": 0.0,
      "MaxTokens": 4096
    }
  }
}

With configuration binding, no code options are needed:

builder.Services.AddRevealAI()
    .AddOpenAI();

Options

Property	Type	Default	Description
ApiKey	string	""	Required. Your OpenAI API key.
Model	string	"gpt-4.1"	The model to use for chat completions.
Temperature	float?	0.0	Controls randomness (0.0 to 2.0). Lower values are more deterministic.
MaxTokens	int?	4096	Maximum number of tokens to generate in the response.
TopP	float?	1.0	Nucleus sampling parameter.
Endpoint	string	""	Custom endpoint URL. Use this for OpenAI-compatible APIs.
ReasoningEffort	string?	null	Reasoning effort level for reasoning models (e.g., o3, o4).

Reasoning Models

The provider automatically detects reasoning models (o1, o3, o4, gpt-5) and adjusts behavior accordingly:

• Temperature is disabled for reasoning models
• Reasoning effort can be configured via the ReasoningEffort option

builder.Services.AddRevealAI()
    .AddOpenAI(options =>
    {
        options.ApiKey = builder.Configuration["RevealAI:OpenAI:ApiKey"];
        options.Model = "o3";
        options.ReasoningEffort = "Medium";
    });

Custom Endpoints

If you are using an OpenAI-compatible API (such as a local model server), you can specify a custom endpoint:

builder.Services.AddRevealAI()
    .AddOpenAI(options =>
    {
        options.ApiKey = "your-api-key";
        options.Endpoint = "https://your-custom-endpoint.com/v1";
        options.Model = "your-model-name";
    });

Never Commit API Keys

Never commit API keys to source control. Always use environment variables, User Secrets, or a secure key management service.