skills.rest
villadalmine
villadalmineCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

whisper-stt

Transcribe Telegram voice notes and generate speech replies via in-cluster Whisper and Edge TTS.

1|Updated Apr 3, 2026
Search Tags:#kubernetes#whisper#tts#stt#arm64#telegram#openclaw
One-click install
npx skills add https://github.com/villadalmine/infra --skill whisper-stt-villadalmine
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: whisper-stt
Source: https://github.com/villadalmine/infra/tree/main/skills/whisper-stt
Command: npx skills add https://github.com/villadalmine/infra --skill whisper-stt-villadalmine

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It removes the need for external speech APIs by providing a fully in-cluster audio pipeline, so Telegram voice notes can be transcribed locally and assistant replies can be spoken back without API keys.

Core Features & Use Cases

  • Local speech-to-text: Converts Telegram voice notes into text using an in-cluster Whisper server.
  • Free text-to-speech: Generates spoken replies with Microsoft Edge neural voices without paid services.
  • OpenClaw integration: Connects audio transcription and response audio through existing OpenClaw configuration and deployment defaults.
  • Use case: Enable a Telegram bot to accept voice messages, turn them into prompts for an agent, and reply with either text or voice as needed.

Quick Start

Deploy the Whisper STT service and then install OpenClaw so the Telegram bot can use local transcription and voice replies.

Frequently Asked Questions about whisper-stt

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

FAQPage Schema
How do I transcribe Telegram voice notes locally on Kubernetes without external speech APIs?▼

You can transcribe Telegram voice notes locally on Kubernetes by deploying an internal Whisper server for speech-to-text and configuring OpenClaw to route Telegram audio through this in-cluster pipeline without API keys.

Can I use Microsoft Edge TTS to generate spoken replies for a Telegram bot?▼

Yes, you can use Microsoft Edge neural voices to generate spoken replies for a Telegram bot. This Skill configures outbound text-to-speech audio for OpenClaw deployments without requiring paid TTS services.

Does whisper-stt work on ARM64 Kubernetes nodes?▼

Yes, this local speech-to-text pipeline is designed for ARM64 Kubernetes nodes. It applies to OpenClaw and Hermes deployments needing in-cluster audio transcription and voice-note handling on ARM64 architecture.

What is needed to set up local speech-to-text for OpenClaw?▼

To set up local speech-to-text for OpenClaw, you need an internal Whisper server, an OpenAI-compatible base URL ending in /v1, cache-backed model storage, and Microsoft Edge TTS configuration for outbound audio.

How do I avoid external speech API costs when processing Telegram voice messages?▼

You avoid external speech API costs by using a fully in-cluster audio pipeline. This Skill transcribes Telegram voice notes with a local Whisper server and replies with free Microsoft Edge TTS voices, requiring no external API keys.

Are there limitations to running local Whisper speech-to-text in a Kubernetes cluster?▼

Running local Whisper speech-to-text in Kubernetes requires cache-backed model storage and an OpenAI-compatible base URL ending in /v1. You must also configure Microsoft Edge TTS separately to generate outbound audio replies.

Related Skills

Explore More Software Engineering Skills→
NITISH-gitbitNITISH-gitbit

whisper

Transcribe spoken language to text using OpenAI's Whisper model.

Community
Intermediate
storyclaw-officialstoryclaw-official

mira-voice-f5

Generate Taiwanese female speech offline with F5-TTS and send via Telegram.

Official
Advanced
coolishagentcoolishagent
30

lobstalk

Connect OpenClaw agents to Telegram groups for peer-to-peer conversations.

Community
Intermediate
softweeksoftweek

openai-whisper

Transcribe and translate local audio files using the Whisper CLI.

Community
Intermediate