The /deepgram command enables you to transcribe audio and convert text to speech using AI-powered technology. Perfect for:

• Transcribing audio files
• Converting text to speech
• Analyzing speech patterns
• Managing API keys
• Tracking usage

​

Basic Usage

Use the command to interact with Deepgram:

/deepgram transcribe audio from https://example.com/audio.mp3

/deepgram convert "Hello world" to speech

/deepgram check usage for project 12345-67890-abcdef

​

Key Features

​

Audio Transcription

• Transcribe audio files
• Support multiple languages
• Real-time transcription
• Batch processing
• Custom vocabulary

​

Text-to-Speech

• Convert text to audio
• Multiple voice models
• Various audio formats
• Adjustable sample rates

​

Advanced Features

• Speaker diarization
• Sentiment analysis
• Automatic summarization
• Smart formatting
• Punctuation

​

Project Management

• View project info
• Track usage statistics
• Manage API keys
• Monitor costs

​

Example Commands

​

Basic Transcription

/deepgram transcribe https://example.com/meeting.mp3 using nova-2 model

​

Transcription with Options

/deepgram transcribe audio.mp3 with punctuation, paragraphs, and speaker detection

​

Custom Vocabulary

/deepgram transcribe audio.mp3 with keywords "Deepgram:3, API:2, transcription:2"

​

Text-to-Speech

/deepgram speak "Welcome to our service" using aura-asteria-en voice as mp3

​

Get Project Info

/deepgram list all projects

​

Check Usage

/deepgram show usage for project 12345 from 2023-01-01 to 2023-01-31

​

Create API Key

/deepgram create API key "Production key" with read and write permissions for project 12345

​

List Models

/deepgram show available models

​

Model Options

​

Speech-to-Text Models

• nova-2: Latest, most accurate
• enhanced: Good balance
• base: Fastest, basic accuracy

​

Text-to-Speech Voices

• aura-asteria-en: Natural English voice
• Multiple voice options available

​

Transcription Options

• smart_format: Intelligent formatting
• punctuate: Add punctuation
• paragraphs: Format paragraphs
• utterances: Separate speakers
• diarize: Speaker identification
• sentiment: Emotion analysis
• summarize: Auto-summarization

​

Tips

• Use nova-2 model for best accuracy
• Add keywords for domain-specific terms
• Enable smart_format for readable output
• Check usage regularly to monitor costs