deepgram-dotnet-audio-intelligence

name: deepgram-dotnet-audio-intelligence description: Use when writing or reviewing C# code in this repo that enables Deepgram intelligence overlays on Speech-to-Text requests. Covers PreRecordedSchema analytics flags such as Summarize, Topics, Intents, Sentiment, DetectLanguage, DetectEntities, Diarize, and Redact, plus the smaller live-streaming subset on LiveSchema. Use deepgram-dotnet-speech-to-text for plain transcription and deepgram-dotnet-text-intelligence for analytics on already-transcribed text.

Using Deepgram Audio Intelligence (.NET SDK)

Audio analytics live on top of the same STT clients. Turn features on through schema properties.

When to use this product

• You have audio and want transcript + analytics in one request.
• REST is the main path for summaries, topics, intents, sentiment, and language detection.

Use a different skill when:

• You only need transcript text → deepgram-dotnet-speech-to-text.
• You already have text and want /read → deepgram-dotnet-text-intelligence.

Authentication

dotnet add package Deepgram

using Deepgram;

Library.Initialize();
var client = ClientFactory.CreateListenRESTClient();

ClientFactory reads credentials from the DEEPGRAM_API_KEY (or DEEPGRAM_ACCESS_TOKEN) environment variable by default. To pass them explicitly, use ClientFactory.CreateListenRESTClient(apiKey: "...", options: ...) / CreateListenWebSocketClient(apiKey: "...", options: ...). DeepgramHttpClientOptions / DeepgramWsClientOptions throw if neither the env var nor an explicit credential is provided.

Feature availability: REST vs WebSocket

Quick start — REST

using Deepgram.Models.Listen.v1.REST;

var client = ClientFactory.CreateListenRESTClient();

var response = await client.TranscribeUrl(
    new UrlSource("https://dpgr.am/spacewalk.wav"),
    new PreRecordedSchema()
    {
        Model = "nova-3",
        Punctuate = true,
        SmartFormat = true,
        Diarize = true,
        Summarize = "v2",
        Topics = true,
        Intents = true,
        Sentiment = true,
        DetectLanguage = true,
        DetectEntities = true,
        Redact = new List<string> { "pii", "numbers" },
        Language = "en",
    });

Console.WriteLine(response.Results.Channels[0].Alternatives[0].Transcript);
Console.WriteLine(response.Results.Summary);

Quick start — live subset

using Deepgram.Models.Listen.v2.WebSocket;

var liveClient = ClientFactory.CreateListenWebSocketClient();

await liveClient.Subscribe(new EventHandler<ResultResponse>((sender, e) =>
{
    Console.WriteLine(e.Channel.Alternatives[0].Transcript);
}));

await liveClient.Connect(new LiveSchema()
{
    Model = "nova-3",
    Diarize = true,
    Redact = new List<string> { "pii" },
    Encoding = "linear16",
    SampleRate = 16000,
});

Key params

REST (PreRecordedSchema): Summarize, Topics, Intents, Sentiment, DetectLanguage, DetectEntities, CustomTopic, CustomTopicMode, CustomIntent, CustomIntentMode, Diarize, DiarizeVersion, DiarizeModel, Redact, Utterances, plus the regular STT knobs.

Live (LiveSchema): Diarize, Redact, UtteranceEnd, VadEvents, Punctuate, SmartFormat, plus normal streaming STT settings.

API reference (layered)

1. In-repo source of truth:
   • Deepgram/Models/Listen/v1/REST/PreRecordedSchema.cs
   • Deepgram/Models/Listen/v2/WebSocket/LiveSchema.cs
   • Deepgram/Clients/Listen/v1/REST/Client.cs
   • Deepgram/Clients/Listen/v2/WebSocket/Client.cs
2. Canonical OpenAPI (REST): https://developers.deepgram.com/openapi.yaml
3. Canonical AsyncAPI (streaming / WSS): https://developers.deepgram.com/asyncapi.yaml
4. Context7:
   • repo mirror: https://context7.com/deepgram/deepgram-dotnet-sdk
   • docs corpus: /llmstxt/developers_deepgram_llms_txt
5. Product docs:
   • https://developers.deepgram.com/docs/stt-intelligence-feature-overview
   • https://developers.deepgram.com/docs/summarization
   • https://developers.deepgram.com/docs/topic-detection
   • https://developers.deepgram.com/docs/intent-recognition
   • https://developers.deepgram.com/docs/sentiment-analysis
   • https://developers.deepgram.com/docs/language-detection
   • https://developers.deepgram.com/docs/redaction
   • https://developers.deepgram.com/docs/diarization

Gotchas

1. Summarize on STT is a string. Use Summarize = "v2", not true.
2. Most intelligence overlays are REST-only. The live schema does not expose summary/topic/intent/sentiment/language options.
3. Redaction is a List<string>. This SDK does not expose the Python-style union of string-or-array.
4. Custom topics and intents need modes. Set CustomTopicMode / CustomIntentMode or the model behavior will not match expectations.
5. Diarization is available in both paths, but the event shapes differ. Live results expose speaker data word-by-word in streaming responses.

Example files in this repo

• examples/speech-to-text/rest/summary/Program.cs
• examples/speech-to-text/rest/sentiment/Program.cs
• examples/speech-to-text/rest/topic/Program.cs
• examples/speech-to-text/rest/intent/Program.cs
• examples/speech-to-text/rest/file/Program.cs

Central product skills

For cross-language Deepgram product knowledge — the consolidated API reference, documentation finder, focused runnable recipes, third-party integration examples, and MCP setup — install the central skills:

npx skills add deepgram/skills

This SDK ships language-idiomatic code skills; deepgram/skills ships cross-language product knowledge (see api, docs, recipes, examples, starters, setup-mcp).