n8n speech-to-text automation with ParakeetASR

Transform audio files into text automatically using NVIDIA's Parakeet model and an n8n workflow. Originally published on Medium.

Why ParakeetASR beats the giants

Most developers assume faster = more expensive. With ParakeetASR you get both:

• 12× cheaper than OpenAI Whisper ($0.06 vs $0.36 per hour)
• 3380× faster processing (1 second vs 56+ minutes for 1 hour of audio)
• Simple pricing: 1 credit = 1 minute, no complex billing tiers
• True free tier: 100 minutes monthly, no strings attached
• No vendor lock-in: standard REST API, works with any platform

Real cost comparison — transcribing 10 hours monthly:

| Service | Cost |

|---|---|

| ParakeetASR | $0.60 (or FREE on Basic) |

| OpenAI Whisper | $3.60 |

| AWS Transcribe | $6.00 |

| Google Speech-to-Text | $7.20 |

What we're building

A complete speech-to-text automation using:

• ParakeetASR API — ultra-fast GPU-powered transcription
• n8n — open-source workflow automation
• Simple web form — easy audio file uploads

By the end you have a transcription system that processes 12+ audio formats and returns clean text.

Step 1: Get your RapidAPI key

1. Create an account at rapidapi.com
2. Search for ParakeetASR
3. Subscribe to a plan (Basic has a free tier)
4. Copy the X-RapidAPI-Key from the Transcribe tab

Step 2: Import the workflow

The workflow is three nodes wired in a line:

1. Form Trigger (formTrigger) — a web form with a single required audio file field
2. HTTP Request — POST to https://parakeetasr-transcription-api.p.rapidapi.com/transcribe with headers x-rapidapi-host and x-rapidapi-key, body as multipart/form-data with the uploaded file bound as file
3. Code — join the response segments into one transcript:

// Grab the API response
const data = $input.first().json;

// Build the transcript (each segment on its own line)
const transcript = data.segments
  .map(seg => seg.segment)
  .join('\n');

return [
  {
    json: {
      transcript,
    },
  },
];

To import: create a new workflow → ⋯ menu → Import from file → paste the JSON → Import.

Step 3: Configure your API key

Replace YOUR_API_KEY in the HTTP Request node headers with your real RapidAPI key, then save.

Step 4: Test the workflow

1. Activate the workflow
2. Copy the form URL from the Form Trigger node
3. Open it in a browser, upload an audio file (MP3, WAV, …), submit
4. The transcript comes back as text

Who it's for

• Content creators — YouTube subtitles, podcast show notes, near real-time captioning
• Business — meeting transcripts, support-call analytics, voice memos
• Accessibility & education — captioning, voice interfaces, lecture transcription