LigthsON - Voice Agent API Sandbox & Dashboard

🎙️ Voice & Text Transaction Tester

Speak naturally or type a financial transaction sentence to test STT, NLU entity extraction, and Firestore CRUD payload generation.

Integration Model:

Model A (JSON Schema Only) Model B (Direct Firestore CRUD)

OR TYPE UTTERANCE

📊 Extracted API Response

Confidence: --

🔊 Result will appear here after processing...

Transaction Type --

Amount --

Category --

Date --

RAW JSON API SCHEMA (Section 3.2 Specification)

// JSON output will be rendered here

🔌 Integration Code Guide (How to connect React Native / Backend to this hosted API)

Use the endpoints below to connect your mobile application or backend to this hosted Voice Agent service:

POST /api/voice/process-text Process natural language text utterance or multi-turn conversational session

POST /api/voice/process-audio Upload recorded audio file (.wav/.m4a/.mp3) or base64 audio stream for STT + NLU

POST /api/voice/parse Model A Parse-Only endpoint (returns JSON schema without DB write)

cURL Example:

curl -X POST https://your-hosted-domain.com/api/voice/process-text \
  -H "Content-Type: application/json" \
  -d '{"text": "Spent 500 rupees on groceries today", "model": "A"}'