Labs

What we're building next.

VoxTN Labs is where we build what doesn't exist yet — sovereign AI infrastructure, voice intelligence, and clinical communication systems designed for Canada's most underserved communities.

Periyanna AI

Voice-first Tamil companion for Canadian diaspora seniors.

A mobile companion app pairing low-latency voice interaction with sovereign data handling. Phase 1 ships a mobile companion; Phase 2 adds radio injection; Phase 3 integrates across the VoxTN ecosystem.

  • Filler-first sub-800ms response architecture
  • Diglossia-aware dual-transcript for natural Tanglish
  • Sovereign PII proxy preserving Canadian data residency

AGIL Clinical Bridge

Multilingual accessibility and clinical communication for Canadian healthcare.

A privacy-preserving interpretation and care-summary platform built on five non-negotiable pillars: persona fidelity, sovereign data control, low-latency voice, deterministic orchestration, and shared ecosystem intelligence. PIPEDA-compliant by design.

  • Real-time English ↔ Tamil interpretation in clinical settings
  • Patient-friendly summaries reviewed by clinicians
  • Bounded emergency profile access for first responders

TrueNorth AI / VoxSync

Sovereign Canadian AI platform for regulated industries.

Multi-model orchestration spanning Sarvam 30B (Tamil), DeepSeek-R1-70B (reasoning), Gemma 4 (local), and Vertex AI Montreal (cloud bridge). Includes a sovereignty dashboard with live data-residency status and an intelligence-routing toggle.

  • Self-hosted models in Toronto
  • Live data-residency map and audit trail
  • CRA NETFILE certification pathway documented

Aravam Streaming Media Platform

Full-stack Tamil media streaming infrastructure.

Multi-station radio and streaming infrastructure with archive pipelines, transcript indexing, and AI radio bulletin generation from approved community sources. Periyanna AI voice ducks into broadcasts as overlay.

  • Multi-station architecture
  • Transcript indexing and archive pipeline
  • AI bulletin generation from approved sources

Screen Motion

UI-driven SaaS product video engine.

Deterministic, CPU-first scene rendering using real product UI components rather than screen recording. Pipeline composes Playwright scenes and FFmpeg output, with audio synced via Aravam TTS.

  • Real product UI, not screen capture
  • Deterministic scene rendering
  • Tamil and English voiceover via Aravam

Nova Motion

Script-to-video orchestration engine — on the roadmap.

A standalone GPU-accelerated video generation engine taking script → voice → asset pipeline → composition → MP4. Three-tier asset fallback: stock API → Ken Burns still image → generative.

  • Tamil-first voiceover via Aravam TTS
  • Three-tier asset fallback strategy
  • Composable script-to-video pipeline

Ad-Genius

AI-assisted ad creation with print-safe magazine output.

A 10-phase implementation plan covering campaign management, brand-memory pgvector store, and CMYK PDF export for print workflows like Thaiveedu.

  • Print-safe magazine output (CMYK, safe zones, trim-aware)
  • Campaign engine with affiliate/API/whitelabel hooks
  • Future Nova Motion video ad integration

BrightSpeak / Project English

Educational initiative connecting volunteers with underserved students.

A platform pairing Canadian university volunteers with students in rural Sri Lanka for live English instruction (Grades 7–12), with a volunteer scheduling system and certification pathway. Vision: expand globally.

  • Live instruction via Zoom and streaming platforms
  • Volunteer scheduling and certification
  • "One English lesson at a time"

VoxTN Sovereign Intelligence Research Archive

47-shelf research archive across 6 domains.

Internal research base classifying everything from server and OS infrastructure to AI models, voice and audio stack, sovereign platform architecture, Canadian compliance, and the multi-phase scaling roadmap.

  • 6 domains, 23 sub-categories
  • 47 shelves classified April 2026
  • Foundation for sovereign platform decisions

Want to collaborate?

If your organization works in any of these areas — healthcare, education, sovereign AI — we'd like to hear from you.