Voxa.Services.OpenAIRealtime 0.7.2-alpha

This is a prerelease version of Voxa.Services.OpenAIRealtime.

dotnet add package Voxa.Services.OpenAIRealtime --version 0.7.2-alpha

NuGet\Install-Package Voxa.Services.OpenAIRealtime -Version 0.7.2-alpha

This command is intended to be used within the Package Manager Console in Visual Studio, as it uses the NuGet module's version of Install-Package.

<PackageReference Include="Voxa.Services.OpenAIRealtime" Version="0.7.2-alpha" />

For projects that support PackageReference, copy this XML node into the project file to reference the package.

<PackageVersion Include="Voxa.Services.OpenAIRealtime" Version="0.7.2-alpha" />
                    

                            Directory.Packages.props

<PackageReference Include="Voxa.Services.OpenAIRealtime" />
                    

                            Project file

For projects that support Central Package Management (CPM), copy this XML node into the solution Directory.Packages.props file to version the package.

paket add Voxa.Services.OpenAIRealtime --version 0.7.2-alpha

The NuGet Team does not provide support for this client. Please contact its maintainers for support.

#r "nuget: Voxa.Services.OpenAIRealtime, 0.7.2-alpha"

#r directive can be used in F# Interactive and Polyglot Notebooks. Copy this into the interactive tool or source code of the script to reference the package.

#:package Voxa.Services.OpenAIRealtime@0.7.2-alpha

#:package directive can be used in C# file-based apps starting in .NET 10 preview 4. Copy this into a .cs file before any lines of code to reference the package.

#addin nuget:?package=Voxa.Services.OpenAIRealtime&version=0.7.2-alpha&prerelease
                    

                            Install as a Cake Addin

#tool nuget:?package=Voxa.Services.OpenAIRealtime&version=0.7.2-alpha&prerelease
                    

                            Install as a Cake Tool

The NuGet Team does not provide support for this client. Please contact its maintainers for support.

Voxa.Services.OpenAIRealtime

OpenAIRealtimeProcessor — a single drop-in FrameProcessor that bridges your Voxa pipeline to the OpenAI Realtime API. Full-duplex streaming, server-side voice activity detection, native interruption, sub-400 ms turn-taking. The C# equivalent of Pipecat's OpenAIRealtimeBetaService.

Why use this instead of Whisper + TTS chained?

Chaining Whisper REST → LLM → TTS REST accumulates 1.3–1.9 s of latency per turn and Whisper hallucinates words from breath/silence. The Realtime API solves both at once:

	`Whisper REST + TTS REST` chain	`OpenAIRealtimeProcessor`
Time to first audio	1.3–1.9 s	250–400 ms
Hallucinations on silence	Frequent	None — server VAD gates speech
Interruption	Manual gate, easy to mistime	Native, server-driven
API calls per turn	2× HTTP	1× WebSocket session

Install

dotnet add package Voxa.Services.OpenAIRealtime

Quickstart

using Voxa.Pipelines;
using Voxa.Services.OpenAIRealtime;
using Voxa.Transports.WebSocket;

var pipeline = Pipeline.Build()
    .Source(new WebSocketAudioSource(clientWs))
    .Then(new OpenAIRealtimeProcessor(new OpenAIRealtimeOptions
    {
        ApiKey = builder.Configuration["OpenAI:ApiKey"]!,
        Model = "gpt-realtime-mini",
        Voice = "alloy",
        Instructions = "You are a friendly voice assistant. Keep responses brief.",
    }))
    .Sink(new WebSocketAudioSink(clientWs));

await using var runner = new PipelineRunner(pipeline);
await runner.StartAsync();
await runner.WaitAsync();

That's the whole thing. No SilenceGate, no separate STT, no separate TTS — the Realtime API handles all of it server-side.

Configuration

Option	Default	Notes
`ApiKey`	(required)	Sent as `Authorization: Bearer <key>`.
`Endpoint`	`wss://api.openai.com/v1/realtime`	Override only when proxying. The model is appended as `?model=...` automatically.
`Model`	`gpt-realtime-mini`	Or `gpt-realtime`, `gpt-4o-realtime-preview`.
`Voice`	`alloy`	One of `alloy`, `ash`, `ballad`, `coral`, `echo`, `sage`, `shimmer`, `verse`.
`Instructions`	`null`	System prompt.
`TurnDetection.Threshold`	`0.5`	Server-VAD activation probability. Lower = more sensitive.
`TurnDetection.PrefixPaddingMs`	`300`	Audio prepended to detected speech for context.
`TurnDetection.SilenceDurationMs`	`500`	Sustained silence before turn-end.
`Tools`	`[]`	Function tools the model can invoke. Wire results back via `ToolCallResultFrame`.
`InputSampleRate` / `OutputSampleRate`	`24000`	The Realtime API expects pcm16 @ 24 kHz.

Tool calling

The processor emits a ToolCallRequestFrame when the model wants to invoke a tool. Your downstream processor handles it and pushes back a ToolCallResultFrame (e.g. via the MicrosoftAgentsProcessor adapter or your own dispatcher) — the processor forwards that back into the session and asks the model to continue.

Auth note

The transport sends both Authorization: Bearer <ApiKey> and OpenAI-Beta: realtime=v1. If you proxy through a gateway that strips the OpenAI-Beta header, the connection will fail.

License

MIT.

Product	Compatible and additional computed target framework versions.
.NET	net10.0 is compatible. net10.0-android was computed. net10.0-browser was computed. net10.0-ios was computed. net10.0-maccatalyst was computed. net10.0-macos was computed. net10.0-tvos was computed. net10.0-windows was computed.

Compatible target framework(s)

Included target framework(s) (in package)

Learn more about Target Frameworks and .NET Standard.

net10.0
- Microsoft.Extensions.Logging.Abstractions (>= 10.0.7)
- Voxa.Core (>= 0.7.2-alpha)

NuGet packages

This package is not used by any NuGet packages.

GitHub repositories

This package is not used by any popular GitHub repositories.

Version	Downloads	Last Updated
0.7.2-alpha	86	7/10/2026
0.7.1-alpha	71	7/10/2026
0.7.0-alpha	70	7/10/2026
0.6.0-alpha	85	6/22/2026
0.5.0-alpha	70	6/13/2026
0.4.0-alpha	1,125	5/10/2026