speech-to-text

Transcribe and diarize spoken audio into text using the Saaras model.

Updated Jun 18, 2026

Search Tags:#api #voice #transcription #multilingual #websocket #speech-to-text #diarization

Ready to run

npx skills add https://github.com/abhishekmmgn/didactic-invention --skill speech-to-text-abhishekmmgn

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: speech-to-text
Source: https://github.com/abhishekmmgn/didactic-invention/tree/main/.agents/skills/speech-to-text
Command: npx skills add https://github.com/abhishekmmgn/didactic-invention --skill speech-to-text-abhishekmmgn

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Transcribing spoken audio into accurate text across multiple languages for searchable transcripts and voice-enabled apps.

Core Features & Use Cases

Real-time & batch Transcription: convert live speech or long recordings to text with optional diarization.
Language & Output Flexibility: auto language detection and five output modes (transcribe, translate, verbatim, translit, codemix).
Use Case: Build a multilingual voice assistant that transcribes meetings and generates translated summaries.

Quick Start

Create a SarvamAI client with your API key and call the speech_to_text.transcribe method with your audio file.

Frequently Asked Questions about speech-to-text

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

FAQPage Schema

How do I transcribe real-time streaming audio to text across multiple languages?▼

You can transcribe real-time streaming audio to text across 23 languages using the Saaras model via WebSocket. It supports automatic language detection and multiple output modes for live voice-enabled applications.

Can I use speech-to-text transcription for batch processing long recordings?▼

Speech-to-text transcription supports batch processing for long recordings with optional diarization. You can send audio files to the Saaras model API to generate searchable multilingual transcripts.

Does the transcription API support automatic language detection and translation?▼

Yes, the transcription API provides automatic language detection across 23 languages and includes five output modes: transcribe, translate, verbatim, translit, and codemix for flexible text conversion.

What do I need to set up multilingual speech-to-text transcription in Python?▼

You need a SarvamAI API subscription key and the base URL https://api.sarvam.ai/v1. Create a SarvamAI client in Python and call the speech_to_text.transcribe method with your audio file.

How does diarization work when transcribing spoken audio into text?▼

Diarization separates and labels different speakers in spoken audio during transcription. The Saaras model applies diarization to both real-time streaming and batch audio processing to identify individual speakers.

What is the best way to build a multilingual voice assistant with meeting transcription?▼

Use the Saaras speech-to-text model to transcribe meetings across 23 languages with automatic detection, then apply the translate output mode to generate translated multilingual summaries for voice-enabled applications.