iTantra -Indian Multilingual TTS & STT Aided Neural Transceiver Radio Access for low bitrate links
“ISRO needs a smart Android app that works like an offline walkie-talkie for remote or disaster areas where internet is absent. Instead of sending heavy raw voice recordings, the app instantly listens to speech in 10 Indian languages, converts it to compact text, transmits it via Bluetooth or Wi-Fi to another phone, and reads it out loudly as clear audio.”
Deconstructed Problem Blueprint
WHAT TO BUILD?
An offline-first Android application acting as a neural transceiver (walkie-talkie and messenger) integrating lightweight Speech-to-Text (STT) and Text-to-Speech (TTS) models optimized for 10 Indian languages.
WHO IS IT FOR?
Field personnel, disaster response teams, remote villagers, and citizens in low-connectivity or emergency regions relying on the Department of Space and ISRO infrastructure.
WHY DOES IT EXIST?
Raw audio files are too data-heavy to transmit over low-bandwidth or ad-hoc wireless emergency links. Furthermore, voice communication is vital for inclusivity since emergency alerts must cater to illiterate or distressed individuals without relying on cloud servers.
INPUT → PROCESS → OUTPUT ARCHITECTURE PIPELINE
Microphone voice input from the user in 10 regional Indian languages; incoming text/data packets via Wi-Fi or Bluetooth from another device.
Local pause detection and silence trimming, on-device lightweight STT model processing, local packetization of text, wireless transmission via local Wi-Fi/Bluetooth socket, on-device TTS synthesis, and priority high-volume alert triggering.
Transmitted text packets over local connection, synthesized voice audio played back via speaker, and non-interruptible high-volume emergency alerts.
CONSTRAINTS
- Must run 100% offline without any internet-hosted APIs or cloud services
- Must support 10 specific Indian languages (Hindi, Gujarati, Marathi, Kannada, Malayalam, Tamil, Telugu, Odia, Bengali, English)
- Low footprint: Optimized for low and mid-range Android mobile phones regarding RAM, flash memory, and idle CPU usage
- Open-source constraint: Strictly no proprietary or commercial voice SDKs allowed
JURY METRICS
- Efficiency: App size, RAM usage, and battery/CPU footprint during idle listening
- Accuracy: Low Word Error Rate (WER) for STT and high human legibility/flow for TTS
- Latency: End-to-end delay from words spoken to audio playback on the receiving phone, along with Real-Time Factor (RTF)
CORE BOTTLENECK
Achieving low-latency, high-accuracy multilingual STT and natural TTS execution completely locally on low-end Android hardware without crashing or draining the battery.
Why You Should Choose This PS (Pros)
- Extremely high social impact addressing national disaster management and remote communication
- Clear, modular technical deliverables (STT module, TTS module, local transport layer)
- Direct alignment with ISRO's mission objectives for inclusive remote communication
Challenges & Risks to Watch (Cons)
- High AI model optimization complexity to fit 10 languages into a lightweight mobile package
- Handling noisy, real-world acoustic environments during disaster scenarios
- Managing cross-device local socket synchronization over Bluetooth/Wi-Fi mesh
Recommended Tech Stack & Grand Finale Features
Killer Features to Impress the SIH Jury
View Official SIH 2026 Portal Raw Text▼
Background As vocal audio information is very data intensive making it difficult to transmit through low data rate links. In alert and distress based scenarios Transmitting Audio information is critical instead of written message as it will be more inclusive and will cater to everyone even if they are literate or not.
Description Build an Android App with lightweight, highly accurate STT and TTS models for 10 Indian Languages (Hindi, Gujarati, Marathi, Kannada, Malayalam, Tamil, Telugu, Odia, Bengali, English) that runs locally on a low-power device. The system’s STT module when activated after detecting pauses and stoppages should form the sentences detected and must instantly and efficiently stream the data through wifi/Bluetooth connected embedded device or another phone with same application with minimal latency. The systems TTS module when activated after receiving the Text data should convert it into intelligible speech which will be played as a voice note and alert type messages will be announced at highest volume non-interruptible. To verify the complete loop two phones with same app one in TTS mode and another in STT mode can be connected via wifi or Bluetooth and it should work like a walkie talkie using push to talk feature, if turned off it should work like a phone.
Key Metrics for Evaluation
Software & Framework Restrictions
Expected Solution Teams are expected to deliver a robust, deployable system architecture. A successful submission must strictly satisfy the following technical boundaries: