Shunya LabsShunya LabsPlayground

Horizon Indic · Sanskrit

Sanskrit Speech to Text for Real-World Conversations

Turn spoken Sanskrit into searchable text for recitations, archives and language-learning audio. Shunya Labs Horizon Indic is designed for speech in Indian languages and regional varieties, with a platform for both uploaded recordings and live audio streams.

Horizon Indic Indian languages and regional varietiesBatch uploaded recordingsStreaming live audiosa language route
Sanskrit audio in
Sanskrit audioHorizon IndicSanskrit text
Searchable text for recitations, archives and language-learning audio
Language - SanskritModel - Horizon IndicCode - saScript - Devanagari and other traditional scripts

Language-aware speech recognition

Why Sanskrit Needs a Language-AwareSpeech Recognition Workflow

Sanskrit is used in education, research and cultural institutions. A single clean sample rarely captures the practical demandsof a deployed speech system. Real-world audio includes different microphones, spontaneous phrasing,local place names, names of people, interruptions and background noise.

Real-world audio includes

Microphones, phrasing, names, and noise

  • Different microphones
  • Spontaneous phrasing
  • Local place names
  • Names of people
  • Interruptions
  • Background noise

These details matter when a transcript feeds

Sanskrit audioTranscriptSupport ticket / Community archive / Voice assistant

For Sanskrit, pay particular attention to

Careful pronunciation, compounds, and specialised vocabulary

  • Careful pronunciation
  • Compounds
  • Specialised vocabulary

Select a reference script before testing, and compare the model output against transcripts checked by speakers familiar with the intended audience. A language name in a catalogue is the starting point for evaluation, not a guarantee that every dialect, script or audio channel performs identically.

Horizon Indic

How Sanskrit Speech-to-Text Workswith Horizon Indic

Step 1 · Audio

Supply Sanskrit audio

Supply Sanskrit audio through a supported file upload for batch transcription, or use a streaming connection for live experiences.

Step 2 · Model

Lock the language route

Choose the deployed Indic model and, when supported, lock the correct language route; the publicly referenced code is “sa” (confirm the code with the production language endpoint).

Step 3 · Text

Receive text

Receive text to review, search, display in an application, or pass to an authorised downstream workflow.

Step 4 · Features

Configure batch features

For supported batch configurations, consider diarization, timestamps and optional conversation intelligence. Feature availability can depend on configuration.

Basic flow

Sanskrit audioHorizon IndicTextReview / Search / Application / Downstream workflow

Try it yourself

Test your Sanskrit recordings

Working with Sanskrit audio requires attention to careful pronunciation, compounds, and specialised vocabulary. Test your recordings,review the transcript in the writing system your audience expects, and connect the result to the workflows that matter to your team.

Test real recordings

Interviews, calls, podcasts, and field audio

  • Interviews
  • Phone recordings
  • Podcasts
  • Lectures
  • Field recordings

Review the transcript in

The writing system your audience expects

  • Devanagari
  • Other traditional scripts

Batch and streaming

Batch vs Real-Time Sanskrit Transcription

Batch transcription

When you already have the recording

Use batch transcription when you already have a Sanskrit interview, phone recording, podcast, lecture or field recording. The system processes the completed audio and returns text that reviewers can correct and tag. This mode is a useful starting point for archives and back-office processing because reviewers can inspect the entire recording.

  • Interview
  • Phone recording
  • Podcast
  • Lecture
  • Field recording

Real-time recognition

When the application needs words during a conversation

Use real-time recognition when an application needs the user’s Sanskrit words during a conversation - for example a live caption screen, a voice-driven form or agent assistance. Shunya documents WebSocket streaming in its Horizon platform.

  • Live caption screen
  • Voice-driven form
  • Agent assistance

Validate using the deployed service

Rather than assuming an aggregate benchmark applies to Sanskrit

  • Latency
  • Audio quality
  • Per-language recognition

Writing system

Writing System and SanskritTranscript Quality

Sanskrit is commonly associated with Devanagari and other traditional scripts. Writing-system choices can still vary bycommunity, context or downstream application. Transcription captures the spoken language as text; transliteration changes thescript; translation expresses the meaning in another language. These are distinct operations and should be tested separately.

Commonly associated with

Devanagari and other traditional scripts

  • Devanagari
  • Other traditional scripts

Transcription

Captures the spoken language as text.

Transliteration

Changes the script.

Translation

Expresses the meaning in another language.

Recitation quality depends on pronunciation conventions, compound segmentation and the difference between chanting and conversational speech.

Horizon + Parallax

Sanskrit Translation: Connect Audioto Multilingual Workflows

A typical Sanskrit speech-translation workflow has two clear stages: Horizon transcribes the audio into Sanskrit text, then Parallaxtranslates that transcript into the supported target language. The two outputs can be checked separately. This helps teamsinvestigate whether a problem came from speech recognition, translation, an unexpected script or unclear source audio.

Stage 1 · Horizon

Transcribes the audio into Sanskrit text

Stage 2 · Parallax

Translates that transcript into the supported target language

Sanskrit-to-English workflow

Sanskrit audioHorizon transcriptParallax translationHuman review or business application
Published Parallax evaluationSanskrit
BLEU34
Evaluation tierGood published evaluation tier (still requires testing)

This is a model-level evaluation, not an accuracy guarantee for your application. Validate actual Sanskrit text in your required translation direction.

Validate especially

Proper nouns, terms, numerals, and informal speech

  • Proper nouns
  • Industry-specific terms
  • Numerals
  • Informal speech

For an English-to-Sanskrit experience, check output language-token and script settings first. Translation and code-switch support for a specific pair must be confirmed on the configured model.

Use cases

Practical Sanskrit Use Cases

Regional service teams

Record customer requests in the language speakers use naturally, rather than forcing a Hindi-only or English-only form. The immediate opportunity for Sanskrit is recitations, archives and language-learning audio.

Media and research teams

Create a searchable archive from interviews, field recordings, radio material and other local-language audio. This is especially useful when the source material contains careful pronunciation, compounds, and specialised vocabulary.

Education and archives

Preserve oral lessons, interviews and local histories in a machine-readable form for human review. Keep human review in the workflow until your acceptance thresholds have been met.

Example

Example: From Sanskrit Audioto Actionable Text

An academic team indexes Sanskrit recitations with speaker-checked text and time markers.

Input

A consented Sanskrit recording

A consented Sanskrit recording containing a real enquiry, interview or instruction.

ASR

Horizon Indic and a Sanskrit reviewer

Select Horizon Indic and review the transcript with a Sanskrit speaker, checking names, dates and the expected script.

Optional translation

Parallax, once the pair is validated

Send the reviewed transcript to Parallax only if the exact language pair has been validated.

Output

A searchable record

A searchable record for a CRM, media archive, support dashboard, content workflow or educational resource.

Illustrative workflow

InputASROptional translationOutput

Illustrative workflow only: no generated native-language sentence or translation here is presented as a measured product result. Replace with a tested audio-and-text sample before publishing.

Developers

Developer Integration Notes

Use the Horizon ASR API for speech recognition and Parallax for text translation. For Sanskrit, the editorial language identifier is sa; obtain the accepted code and model ID from the production API before putting an executable snippet on this page. Shunya’s current main documentation calls the Indic model horizon-indic, while older documentation may refer to zero-indic.

sanskrit-batch.txt

Sanskrit audio file upload
horizon-indic · language route: sa
Text to review, search, or display

Recommended configuration checklist

Authentication, audio, model, language, and script

  • Authentication
  • Supported audio format and sampling rate
  • Model ID
  • Explicit language selection versus auto-detect
  • Required script
  • Timestamps
  • Speaker labels
  • Error handling
  • Retention and access rules

Add a tested request/response sample sourced from the current API docs during implementation.

Explore the Horizon ASR API

Questions

Frequently Asked Questions

Can Shunya Labs transcribe Sanskrit audio?

Sanskrit appears in the published Horizon Indic supported-language list. Test the current API with your own speakers, scripts, accents and recording conditions before production use.

Can I transcribe Sanskrit in real time?

Horizon offers both streaming and batch workflows. Verify the exact Sanskrit configuration and performance in the target deployment.

What writing system does Sanskrit speech-to-text use?

Sanskrit is associated with Devanagari and other traditional scripts. Confirm that the deployed output format matches your audience and do not assume every alternate script is automatically available.

Can Sanskrit audio be translated into English?

The normal route is speech transcription followed by Parallax text translation. The published evaluation reports 34 BLEU, but real-world content and direction still require evaluation.

How should I evaluate Sanskrit transcription accuracy?

Use consented recordings that represent your actual speakers and tasks. Review proper nouns, careful pronunciation, compounds, and specialised vocabulary, and output-script consistency rather than relying only on a multilingual platform-wide benchmark.

Horizon Indic

Build Your Sanskrit Voice Workflow

Try real Sanskrit audio in the Shunya Labs playground, inspect the output, and speak with the team about streaming, APIs and translationsupport. Bring the scripts, speaker groups, audio samples and required downstream language that your application will use.