Why use the REST API instead of the Azure Speech SDK?

The REST API is ideal for simple, lightweight applications where you want to avoid the overhead of a large SDK, or when working in restricted environments where installing heavy dependencies is not feasible.

Which audio formats are recommended for best results?

For optimal recognition accuracy, use 16 kHz mono WAV (PCM) files. OGG OPUS is also supported and is useful for reducing file size during transmission.

How does the skill handle different languages?

You can specify the target language using standard BCP-47 codes (e.g., 'en-US', 'fr-FR', 'zh-CN') via the language query parameter to ensure accurate transcription.

Does this skill support real-time streaming transcription?

No, this REST-based implementation is designed for transcribing completed short audio clips. For real-time, continuous streaming, the Azure Speech SDK using WebSockets is required.

What is the maximum audio length supported by this skill?

This specific REST API implementation is limited to audio files of 60 seconds or less. For longer audio files, Microsoft recommends using the Speech SDK or the Batch Transcription API.

Azure Speech-to-Text REST (Python)

Name: Azure Speech-to-Text REST (Python)
Author: microsoft

bymicrosoft

•

1,777

•

Cloud Infrastructure

Integrates Azure's Speech-to-Text REST API into Python workflows for transcribing short audio files without requiring the full Speech SDK.

This Claude Code skill provides a streamlined implementation of the Azure Speech-to-Text REST API for Python, specifically optimized for processing short audio clips under 60 seconds. It enables developers to perform high-accuracy speech recognition using standard HTTP requests, bypassing the complexity and overhead of the full Azure Speech SDK. This is ideal for lightweight applications, voice-command interfaces, and quick transcription tasks where minimal dependencies are preferred. The skill includes support for WAV and OGG formats, chunked transfer for lower latency, and detailed recognition results including confidence scores and profanity filtering.

Key Features

01Supports audio files up to 60 seconds with both simple and detailed response formats

02Lightweight REST-based transcription bypassing the heavy Azure Speech SDK dependency

03Configurable profanity handling and support for over 100+ global languages

04Implementation of chunked transfer encoding for optimized lower-latency processing

051,777 GitHub stars

06Includes both synchronous (requests) and asynchronous (aiohttp) implementation patterns

Use Cases

01Transcribing short voice memos or audio snippets within data processing pipelines

02Building voice-activated command interfaces for lightweight IoT or desktop applications

03Adding speech recognition to serverless functions where small package size is critical

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add microsoft/skills azure-speech-to-text-rest-py

For use in Claude.ai and ChatGPT

Download Skill