Process audio with ElevenLabs via KIE.AI: transcribe, TTS, and isolate audio

This n8n template provides a comprehensive suite of ElevenLabs audio processing capabilities through the KIE.AI API. The workflow includes three independent audio processing workflows: speech-to-text transcription, text-to-speech generation, and audio isolation. Each workflow can be used independently or combined to create complete audio processing pipelines.

Use cases are many: Transcribe audio files to text with speaker diarization, convert text to natural-sounding speech audio, isolate and clean audio by removing background noise, create complete audio processing pipelines from transcription to speech generation, automate podcast transcription and audio enhancement, generate voiceovers from text content, clean up recordings by removing unwanted audio elements, create accessible content by converting text to audio, or process audio files in batch for content creation workflows!

Good to know

- The workflow includes three independent ElevenLabs audio processing capabilities via KIE.AI API: - Speech-to-Text: Transcribes audio to text with speaker diarization and audio event tagging - Text-to-Speech: Converts text to natural-sounding speech with voice customization options - Audio Isolation: Removes background noise and isolates audio sources - Each workflow can be used independently or combined for complete audio processing pipelines - Speech-to-text supports speaker diarization (identifying different speakers) and audio event tagging - Text-to-speech supports multiple voices (Rachel, Adam, Antoni, Arnold, and more) with customizable stability, similarity boost, style, and speed - Audio isolation removes background noise and separates audio sources for cleaner output - KIE.AI pricing: Check current rates at https://kie.ai/ for audio processing costs - Processing time: Varies based on audio length and KIE.AI queue, typically 10-30 seconds for text-to-speech, 30 seconds to 5 minutes for transcription and isolation - Audio requirements: Files must be publicly accessible via URL (HTTPS recommended) - Supported audio formats: MP3, WAV, M4A, FLAC, and other common audio formats - Automatic polling system handles processing status checks and retries for all workflows

How it works

The template includes three independent workflows that can be used separately or combined:

1. Speech-to-Text Transcription: 1. Audio URL Setup: Set the audio file URL in 'Set Audio URL' node 2. Transcription Submission: Audio URL is submitted to KIE.AI API using ElevenLabs speech-to-text model with diarization and event tagging 3. Processing Wait: Workflow waits 5 seconds, then polls the transcription status 4. Status Check: Checks if transcription is complete, queuing, generating, or failed 5. Polling Loop: If still processing, workflow waits and checks again until completion 6. Text Extraction: Once complete, extracts the transcribed text from the API response

2. Text-to-Speech Generation: 1. Text Input Setup: Set the text to convert to speech in 'Set Text Input' node 2. Speech Generation Submission: Text is submitted to KIE.AI API using ElevenLabs text-to-speech multilingual v2 model 3. Processing Wait: Workflow waits 5 seconds, then polls the generation status 4. Status Check: Checks if audio generation is complete, queuing, generating, or failed 5. Polling Loop: If still processing, workflow waits and checks again until completion 6. Audio URL Extraction: Once complete, extracts the generated audio file URL from the API response