ParkVoice: Building a 100% Offline Voice-Enabled OSM Parking Assistant & Turn-by-Turn Road Navigator

Description and use

ParkVoice is a fully offline PWA that lets users find parking spots and receive voice guidance without internet, using local OSM data, a Qwen 3 LLM, and on‑device speech models.

Problem

Drivers lose navigation and parking guidance when cellular connectivity fails, and commercial apps expose location data to telemetry.

Outcome

ParkVoice delivers a fully offline, privacy‑preserving parking assistant and navigation system that eliminates reliance on cloud services and telemetry, enabling reliable use in cellular dead zones.

Workflow

• Download desired OSM region via the Offline Map Manager
• Set origin using GPS, preset, or map tap
• Speak or type a parking query (e.g., "Bring out 13 parking slots nearby Redhills within 4 kms")
• The local Qwen model parses the request and calls offline tools via MCP
• The orchestrator returns parking data and synthesizes a spoken response via Piper TTS
• Select a parking slot and tap Navigate
• The routing engine computes a road route and provides spoken turn‑by‑turn instructions

Evidence

ParkVoice is an end‑to‑end, 100% Offline Progressive Web App (PWA) that lets you find parking spots using natural voice commands, queries a local OpenStreetMap (OSM) spatial database, reasons over constraints using a local Qwen 3 (1.7B GGUF) model via llama.cpp, calculates turn‑by‑turn road navigation routes, and speaks directions aloud with Piper Neural TTS—all running on-device with zero internet, zero cloud API keys, and zero telemetry.

Source

Feature 3: Natural Language Queries with Custom Slots & Radius – Speak or type natural requests without rigid syntax: "Bring out 13 parking slots nearby Redhills within 4 kms" – The local Qwen model extracts your parameters, overrides default limits (upgrades to 13 slots / 4 km), and renders interactive cards + numbered map pins.

Source