Whisper Mapaho

Whisper Mapaho screenshot

Project Description

Whisper Field Report is reporting software for private investigators. You speak your observations, take photos, and it turns them into a structured, timestamped report. Everything runs on the laptop. Nothing leaves the device.

Problems

Surveillance work produces hours of notes and hundreds of photos, and writing the report afterwards takes almost as long as the job. Cloud transcription and AI tools would help, but case material is confidential and often there is no reliable connection in the field anyway.

Solutions

  • On-device voice-to-text with whisper.cpp
  • Automatic, timestamped report entries
  • Photo scene descriptions
  • Object and vehicle detection
  • License plate recognition
  • Export to PDF, ODT and JSON
  • Works fully offline after setup
  • One binary, interface built in

Results

The application is one Go binary with an embedded React interface and a SQLite database. Speech recognition uses whisper.cpp, image description runs moondream2 on llama.cpp, and YOLOv11 detection and plate OCR run through ONNX Runtime. A setup wizard prepares the AI runtime on first run, after which the laptop can go fully offline. Licence keys come from whisper.mapaho.com.

Have a workflow where AI would help but the data cannot leave the building? Talk to us about AI automation.

Information

  • Type : Offline AI Software
  • Year : 2026
  • Our role : Concept, Development, Distribution
  • Website : whisper.mapaho.com

Built with

Go React SQLite whisper.cpp llama.cpp ONNX Runtime YOLOv11