Release Notes: Smart Kiosk Assistant#

2026.2.0#

This release delivers a unified Smart Kiosk Assistant platform with dual React-based user interfaces, OpenVINO-powered AI inference, agentic MCP tool calling, queue-aware ordering, enhanced voice interaction, optional multimodal identity, configurable inference devices, and streamlined deployment.

Key Updates#

  • Dual Kiosk Experiences: Replaced the previous Gradio interface with a React, Vite, and TypeScript application. A single kiosk-ui image supports both operator and customer-facing modes through KIOSK_UI_MODE.

    • Operator mode: Voice/text chat, live queue monitoring, knowledge-base ingestion, audio configuration, and performance dashboards.

    • Customer mode: Queue-aware menu, category navigation, live cart, upsell prompts, queue status, and voice interaction.

  • Agentic Ordering with MCP: Integrated an MCP tool server into kiosk-core to manage catalog browsing, cart operations, order confirmation, and upsell recommendations. Guardrails handle ambiguous item references and invalid quantities.

  • AI Inference with OVMS: Migrated the Qwen3-4B ordering model from in-process OpenVINO inference to OpenVINO Model Server (OVMS), providing a dedicated OpenAI-compatible endpoint for agentic tool calling.

  • Queue-Aware Ordering: Added YOLO-based person counting and RTSP streaming with OpenVINO to monitor queue length, provide a live MJPEG overlay, and dynamically adapt menu recommendations during peak periods.

  • Contextual Upselling: Added a rule-based upsell engine that generates relevant add-on recommendations and surfaces them in both the customer cart and voice responses.

  • Enhanced Speech-to-Text: Upgraded audio-analyzer with OpenAI-compatible SSE streaming transcription, real-time WebSocket transcription, VSS response compatibility, improved multi-speaker segmentation, and persistent speaker enrollment.

  • Speaker Diarization: Enabled speaker attribution across the audio-analyzer and kiosk-core pipeline to improve turn identification during multi-speaker conversations.

  • Improved Text-to-Speech: Added named voice support with faster response generation and more natural speech prosody.

  • Multimodal Identity: Added an optional identity service supporting Face ID and voiceprint authentication using OpenVINO face inference, ECAPA voice embeddings, FAISS indexing, and SQLite-based loyalty profiles. The UI includes login, registration, and enrollment workflows.

  • Flexible Hardware Acceleration: Added per-service inference-device configuration for CPU, GPU, and NPU. NPU passthrough is supported for identity-service, ovms-llm, and audio-analyzer ASR, with independent device configuration for RAG embedding and reranker models.

  • Unified Versioning: All kiosk services and container images are aligned to 2026.2.0:

    • kiosk-core

    • kiosk-ui

    • queue-service

    • identity-service

    • rag-service

    • rtsp-streamer

    • audio-analyzer

    • text-to-speech

    • metrics-collector

  • Simplified Deployment and Operations: Added a Makefile-based workflow for environment initialization, configuration validation, model and sample-video downloads, image builds, service startup, health checks, log monitoring, individual service rebuilds, and cleanup.

  • Sample RTSP Video Provisioning: Added configurable tooling to download and provision sample video clips required by the queue analytics pipeline.

2026.1.0#

The initial release of Smart Kiosk Assistant marks the launch of a voice-enabled interactive application for retail, QSR, Airlines and other customer-facing environments. The application has the following features:

  • Designed as a conversational AI experience, it enables users to engage naturally through speech and receive intelligent, spoken responses in real time.

  • The platform brings together speech recognition, retrieval-augmented generation, and text-to-speech in a seamless, end-to-end voice interaction flow.

  • With browser-based voice capture and natural audio playback, the experience feels intuitive, responsive, and ready for real-world engagement.

  • Smart Kiosk Assistant grounds every response in an ingestible local knowledge base, helping deliver more relevant, context-aware, and business-specific answers.

  • Its integrated AI stack combines kiosk UI, orchestration, speech-to-text, retrieval, and speech synthesis into a unified deployment-ready application.

  • The experience is further enhanced by built-in visibility into model KPIs and live performance data, including runtime model details and latency metrics.

  • Optimized for local and edge deployment, the application leverages OpenVINO acceleration on Intel hardware for efficient AI inference.

  • Docker Compose packaging and flexible configuration make the solution easy to deploy, adapt, and scale across enterprise environments.

  • This launch establishes Smart Kiosk Assistant as a strong foundation for immersive, intelligent, and voice-first digital engagement experiences.