# Deepgram
> Deepgram is a foundational AI company providing the world's fastest and most accurate Speech-to-Text, Text-to-Speech, and Voice Agent APIs. This documentation provides the technical context needed to build voice applications, real-time transcription services, and conversational AI agents.

## Speech-to-Text (ASR)
- [Speech-to-Text Overview](https://developers.deepgram.com/docs/speech-to-text): Introduction to Deepgram's ASR capabilities, offering high-accuracy transcription for both pre-recorded and live audio.
- [Models & Languages](https://developers.deepgram.com/docs/models-languages-overview): A comprehensive reference of available models (including Nova-2 and Nova-3) and supported languages.
- [Pre-Recorded Audio API](https://developers.deepgram.com/docs/pre-recorded-audio): Documentation for the REST API endpoint used to transcribe existing audio files.
- [Live Streaming API](https://developers.deepgram.com/docs/live-streaming-audio): Guide to using WebSockets for real-time, low-latency transcription of streaming audio.
- [Smart Formatting](https://developers.deepgram.com/docs/smart-format): Details on formatting features like punctuation, casing, and paragraph detection.
- [Diarization](https://developers.deepgram.com/docs/diarization): How to detect and label distinct speakers within a single audio stream.

## Text-to-Speech (TTS)
- [Text-to-Speech Overview](https://developers.deepgram.com/docs/text-to-speech): Introduction to Deepgram Aura, a highly responsive TTS model designed for real-time voice agents.
- [TTS API Reference](https://developers.deepgram.com/reference/text-to-speech-api): Technical reference for the `/speak` endpoint to generate human-like speech from text.
- [Streaming TTS](https://developers.deepgram.com/docs/streaming-text-to-speech): Guide to streaming audio output to minimize latency in conversational applications.

## Voice Agents
- [Voice Agent API](https://developers.deepgram.com/docs/voice-agent): Documentation for the Voice Agent API, which orchestrates Speech-to-Text, LLM processing, and Text-to-Speech in a single WebSocket connection.
- [Function Calling](https://developers.deepgram.com/docs/function-calling): How to enable voice agents to perform actions or retrieve data from external APIs during a conversation.
- [Agent Configuration](https://developers.deepgram.com/docs/voice-agent-settings): Settings for customizing agent behavior, including latency controls and interruption handling.

## Audio Intelligence (NLU)
- [Audio Intelligence Overview](https://developers.deepgram.com/docs/audio-intelligence): Suite of NLU features for extracting meaning from audio data.
- [Summarization](https://developers.deepgram.com/docs/summarization): API for automatically generating concise summaries of transcribed conversations.
- [Sentiment Analysis](https://developers.deepgram.com/docs/sentiment-analysis):

Feature for detecting positive, negative, or neutral sentiment in audio segments.
- [Topic Detection](https://developers.deepgram.com/docs/topic-detection): Feature for identifying the primary topics discussed in an audio file.

## SDKs & Developer Tools
- [Deepgram Python SDK](https://developers.deepgram.com/sdks/python): Official Python client library for integrating Deepgram.
- [Deepgram JavaScript SDK](https://developers.deepgram.com/sdks/node): Official Node.js and browser client library.
- [Deepgram Go SDK](https://developers.deepgram.com/sdks/go): Official Go client library.
- [Deepgram .NET SDK](https://developers.deepgram.com/sdks/dotnet): Official C#/.NET client library.
- [API Playground](https://playground.deepgram.com): Interactive sandbox to test models, parameters, and API requests without writing code.

## Management & Reference
- [API Reference](https://developers.deepgram.com/reference): The complete, auto-generated reference for all Deepgram API endpoints.
- [Authentication](https://developers.deepgram.com/docs/authenticating): Guide to creating and managing API keys.
- [Self-Hosted Guide](https://developers.deepgram.com/docs/self-hosted): Documentation for deploying Deepgram models on-premises or in a private VPC.