skills.rest
QuixiAI
QuixiAIOFFICIAL
Explore all skills by this author →
GitHub

audio-analysis

Diarize local audio recordings into timestamped speaker-labeled transcript segments without uploading audio.

609|81|Updated Nov 15, 2024
Search Tags:#transcription#whisper#recording#diarization#local processing#speakers#Audio
One-click install
npx skills add https://github.com/QuixiAI/Hexis --skill audio-analysis
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: audio-analysis
Source: https://github.com/QuixiAI/Hexis/tree/main/skills/installed/audio-analysis
Command: npx skills add https://github.com/QuixiAI/Hexis --skill audio-analysis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Identifying who spoke when in a recording usually requires uploading sensitive audio to a cloud service. This Skill performs speaker diarization entirely on the local device, keeping private conversations private while producing timestamped, speaker-labeled transcripts.

Core Features & Use Cases

  • Local Speaker Diarization: Runs analyze_local_audio on-device with operator approval, never uploading audio to external services.
  • Background Job Management: Starts analysis asynchronously, then polls with action=status until the diarized JSON or SRT artifact is ready.
  • Plain Transcription Fallback: Uses transcribe for simple speech-to-text requests that follow the local/cloud choice in Settings without requiring a diarization job.
  • Use Case: A user has a two-hour meeting recording and a Whisper JSON with timestamps. The Skill starts a local diarization job, polls until completion, and returns a transcript with SPEAKER_* labels the user can map to real names.

Quick Start

Analyze the recording at /recordings/team-sync.wav and tell me who spoke when, using the local diarization job.

Frequently Asked Questions about audio-analysis

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

FAQPage Schema
How do I find out who spoke when in an audio recording?▼

Use local speaker diarization by calling analyze_local_audio with action=start and the exact audio path. The job runs in the background; poll with action=status until it completes, then read the diarized JSON or SRT artifact with SPEAKER_* labels.

How to transcribe audio without speaker diarization?▼

Use the transcribe tool for plain speech-to-text requests. It follows the local/cloud choice in Settings and does not require starting a diarization job, making it the right choice when you only need the text.

Does local audio diarization upload my recordings to the cloud?▼

No. The analyze_local_audio tool runs entirely on the device, requires operator approval before starting, and never uploads audio. Generated artifacts stay in the Hexis cache rather than any external service.

Can speaker diarization detect emotions in a recording?▼

Emotion heuristics are only available when the user explicitly requests them, and the result is a coarse local acoustic estimate. Acoustic heuristics should never be presented as a reliable reading of a speaker's emotional state.

What happens if the audio analysis job fails?▼

When a diarization job fails, the status response reports the cause and the exact next step. Poll with action=status using the same audio_path to retrieve the failure details and follow the reported remediation.

Related Skills

Explore More Content & Communication Skills→
inference-shinference-sh
688

speech-to-text

Transcribe audio files into text with timestamps using Whisper models.

Official
Intermediate
garochee33garochee33
3

transcribe

Transcribe audio and video recordings into text with speaker diarization.

Community
Intermediate
rkz91rkz91
264

pmstudio-meeting-notes

Transforms meeting transcripts into structured notes with extracted actions, decisions, and flagged impacts.

Community
Advanced
tamireslaeao-progtamireslaeao-prog

openai-whisper

Transcribe local audio files to plain text using the Whisper CLI.

Community
Intermediate