Experimental companion

Lumi Transcription

Server-hosted speech recognition for selected OBS sources. Live Twitch delivery remains blocked until the native OBS caption compatibility test passes on the target setup.

<%= providerHealth.healthy ? 'Inference ready' : 'Inference setup required' %>

First usable path

Setup progress

  1. Pair a companionCreate a single-use package for the streaming computer.
  2. Install the OBS bridgeManaged by Lumi Companion; no separate bridge settings.
  3. Install and benchmark a modelSmall English is recommended. Nothing downloads without confirmation.
  4. Select and test a microphoneTest mode must pass before live delivery is enabled.

Lumi host

Inference

Provider
whisper.cpp <%= runtimeManifest.tested_version %>
Worker
<%= providerHealth.state %>
Selected model
<%= providerHealth.model || 'Not loaded' %>
Active sessions
<%= providerHealth.sessions || 0 %>

Lumi AI may already occupy most GPU memory. The MVP warns and benchmarks; it does not unload Lumi AI models automatically.

Access

Paired devices

<%= devices.length %>
<% if (!devices.length) { %>

No companion is paired yet.

<% } %>
    <% devices.forEach((device) => { %>
  • <%= device.name %><%= device.revoked_at ? 'Revoked' : 'Last connected ' + new Date(device.last_connected_at).toLocaleString() %>
    <%= device.id %>
    • <% }) %>

Curated choices

Speech models

<% models.forEach((model) => { %>

<%= model.label %>

<%= model.recommended ? 'Recommended starting point' : 'Fallback option' %> ยท <%= Math.round(model.bytes / 1048576) %> MiB

<%= model.status.valid ? 'Verified' : model.status.installed ? 'Checksum failed' : 'Not installed' %>
<% }) %>

Privacy: raw audio is held only in bounded memory and is never written to disk by default. Diagnostic caption text is retained for seven days unless disabled.