deepgram-transcription

Use when transcribing audio files to text, processing voice recordings, meeting notes, interviews, discovery calls, or any audio-to-text conversion. Also use when working with the Deepgram API for speech recognition.

Model	Source
sonnet	pack: brand

Full Reference

┏━ 🔧 deepgram-transcription ━━━━━━━━━━━━━━━━━━━━┓ ┃ Use when transcribing audio files to text, pro… ┃ ┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛

Deepgram Transcription

Transcribe audio files to text using Deepgram’s Nova-3 speech-to-text REST API. Nova-3 delivers a 47.4% reduction in word error rate for batch processing compared to competitors and supports 60+ languages including real-time multilingual transcription. New accounts get $200 in free credits with no expiration.

Quick Reference

Item	Value
Endpoint	POST https://api.deepgram.com/v1/listen
Auth Header	Authorization: Token YOUR_DEEPGRAM_API_KEY
Env Var	DEEPGRAM_API_KEY (configured in Claude Code settings)
Recommended Model	nova-3 (always specify — default is base-general)
Max File Size	2 GB
Concurrent Requests	50 pre-recorded (Pay-as-you-go / Growth)
Pricing (PAYG)	$0.0077/min Nova-3 mono · $0.0092/min Nova-3 multilingual
Free Credits	$200 on sign-up, no expiration

Reference Index

I want to…	File
Set up API key auth and choose the right model	reference/setup.md
Transcribe a local audio file (curl, Python, Node.js, batch)	reference/transcribe-file.md
Transcribe a remote URL	reference/transcribe-url.md
Handle multi-speaker interviews with diarization	reference/speaker-diarization.md
Enable feature flags (smart_format, diarize, utterances, paragraphs, detect_language, topics, sentiment, intents, summarize, redact)	reference/feature-flags.md
Parse the response JSON and extract transcript data with jq	reference/response-format.md
Handle errors, rate limits, pricing, and retries	reference/error-handling.md

Usage: Read the reference file matching your current task from the index above. Each file is self-contained with code examples and inline gotchas.