For evaluators, content teams, and anyone non-technical

Try it from the dashboard

No SDKs, no terminals, no API keys to wire up. Sign in to the Shunya dashboard, type some text, pick a voice, and hear the output. Same models, same quality, just without the code.

Your journey

Step 1: Sign in

Go to the Shunya Labs dashboard and create an account. You'll land on a workspace home page with three tools in the left rail: ASR, TTS, and Vāķ Translate.

Free tier includes monthly minutes for ASR, TTS, and translation, enough to evaluate.
No credit card required to start.
Workspace is shareable with teammates so a content reviewer and a developer can both work in the same place.

Step 2: Try Speech-to-Text

Drag-and-drop an audio file (WAV, MP3, FLAC, OGG) into the ASR panel. The dashboard returns a transcript with timestamps in seconds. Useful checks:

Recognition accuracy on your real-world audio, accents, noise, code-switching.
Diarization, who spoke when, with per-segment speaker labels.
Emotion and sentiment tags on each segment.
Hinglish handling, mid-sentence Hindi/English mixing without language tagging.

Step 3: Try Text-to-Speech

Type and hear

Paste your text into the input box. Mix Hindi, English, Tamil. The model handles code-switching natively. Hit Generate and the audio plays in your browser.

Compare voices side-by-side

46 voices across 23 Indic languages. Each language has one male and one female speaker, Rajesh / Sunita for Hindi, Varun / Nisha for English, Murugan / Thangam for Tamil, and so on. Click a voice, hit play.

Adjust speed, format, expression

Sliders for speed (0.25× → 4×), volume normalization, and silence trimming. Pick from 11 expression styles, Happy, Sad, News, Narrative, Conversational, and more.

Cross-lingual synthesis

Voice controls accent, not language. A Tamil-native voice can speak English; an English-native voice can read Hindi. Try the same text with different voices to find the right vocal character.

Step 4: Try Vāķ Translation

Paste source text, pick the source and target language from the dropdowns, and translate. 55 Indian languages, 2,970 directional pairs, BLEU 38.5 weighted average. Useful for:

Quickly localizing a script before recording it as TTS.
Spot-checking translation quality on your specific content vocabulary.
Generating multilingual prompts for an IVR or voice agent.

Step 5: Generate an API key (when you're ready)

Once a voice or pipeline feels right, the dashboard hands you off to engineering:

Open API Keys in the dashboard.
Click Create New Key, name it after the environment (e.g. prod-2026-q2).
Copy and store the key securely. It will not be shown again.
Send your developer to High code: API or Small code: SDK.

Who the playground is for

Content & localization teams

Generate voiceovers and audio prompts for IVR, ads, and accessibility, without filing a developer ticket.

Buyers evaluating Shunya

Hear quality before signing a contract. Test on your actual content, your actual languages, your actual voices.

Product managers

Prototype voice features visually before scoping engineering. Click-to-validate replaces back-and-forth specs.

Support and ops

Quickly transcribe a customer call or generate a notification audio without writing code or installing tools.

Ready to integrate?

When the dashboard has answered your "does this work?" questions, hand off to your developer with the Python SDK guide or the REST & WebSocket reference.