LoomVoice · Voice assistant

Talk to AI.
Leave the phone aside.

Children can talk to AI tutors about the textbooks they use at school. Families can ask questions, play media, record notes and set reminders by voice.

In development · A home prototype. Downloads and source code are not yet available.

LoomVoice · Current prototype · Raspberry Pi + speakerphone
Current prototype · Raspberry Pi + speakerphone

See it at work at home.

English narration · 09:53
LoomVoice home voice assistant demo

Watch the demo on YouTube ↗

Author narration uses an AI voice. Real interactions retain their original audio, with English translation subtitles.

01

Learn with the textbooks you use

Follow a child’s question through to the relevant textbook material. See questions, responses and source material together.

02

Shape assistants around different needs

Configure knowledge, roles and voices for different assistants, then select the one you need through wake words and role switching.

03

Connect voice to everyday actions

The recording includes media and TV playback, household records and reminders. After a request, check whether playback starts, a note is saved or an alarm actually rings.

How does a voice request become a response?

01Voice input

Capture and recognize speech

02Assistant

Role and conversation

03Knowledge & tools

Find sources and carry out actions

04Response & result

Spoken response and observable result

An experience overview. Some model and speech capabilities use cloud services.

TODAY

Starting at home.

The current demo runs on a Raspberry Pi and speakerphone. Wake detection, voice interaction and setup are still being refined.

WHAT’S NEXT

Easier setup. More devices.

I want to reduce setup steps and validate more compatible devices. Supported platforms, public installers and source-code release plans are not yet finalized.

FAQ

Can I download it now?

LoomVoice is in development. Public installers and source code are not yet available.

Which devices does it support?

The current demo uses a Raspberry Pi and a speakerphone. More devices and simpler setup are being explored; compatibility is not yet confirmed.

Does everything run locally?

Some model and speech capabilities use cloud services. This is not a fully offline assistant.