audio-processing

Automate audio capture, transcription, and post-processing for dictation and meeting recordings.

1|Updated Mar 26, 2026

Search Tags:#transcription #whisper #pipeline #audio #dictation #stt

Ready to run

npx skills add https://github.com/euroconic/noter --skill audio-processing-euroconic

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: audio-processing
Source: https://github.com/euroconic/noter/tree/main/skills/audio-processing
Command: npx skills add https://github.com/euroconic/noter --skill audio-processing-euroconic

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Building reliable, production-grade audio capture, transcription, and text post-processing for dictation and meeting recordings. This skill defines exact standards and workflows to ensure high-quality ASR results and consistent downstream text output.

Core Features & Use Cases

Enforces precise audio capture constraints (sample rate 16000 Hz, mono, float32) and implements guardrails to avoid low-quality input.
Provides model selection guidance for F1 voice typing and F5 meeting recording, with deterministic transcription settings.
Includes a post-processing pipeline: hallucination filtering, filler-word removal, whitespace normalization, and capitalization to produce clean text.
Supports warmup protocol to prime the ML hardware and a defined latency budget for production readiness.

Quick Start

Configure the audio capture to 16000 Hz, mono, and run the end-to-end STT pipeline to produce post-processed text.

Frequently Asked Questions about audio-processing

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I build a reliable audio transcription pipeline for voice typing and meeting recordings?▼

To build a reliable audio transcription pipeline, you must enforce audio capture constraints like a 16000 Hz sample rate and mono channels, then apply deterministic transcription settings and post-processing to produce clean text.

What audio format constraints are needed for high-quality speech-to-text dictation?▼

High-quality speech-to-text dictation requires audio captured at a 16000 Hz sample rate with mono channels and float32 encoding to ensure production-ready ASR results and avoid low-quality input.

How does post-processing improve Whisper transcription output for meetings?▼

Post-processing improves transcription output by applying hallucination filtering, filler-word removal, whitespace normalization, and capitalization to generate clean, production-ready text from raw ASR results.

Can I use automatic language detection with this audio-to-text pipeline?▼

Yes, the audio-to-text pipeline supports automatic language detection alongside deterministic transcription settings and fp16 operation to ensure consistent, high-quality text output across different languages.

Why does my speech-to-text pipeline have high latency during dictation?▼

High latency during speech-to-text dictation often occurs when skipping the ML hardware warmup protocol or violating the defined latency budget required to prime the system for production readiness.