Documentation

Cross-device state you own

Lux serves reads from the browser and pushes every change across a user's devices over a live WebSocket. One small client, a lightweight Go broker you run on your own cloud account — no Kafka, RabbitMQ, or queue infrastructure.

// local-first read, live push on change
import { featherweight } from '@voxell/featherweight'

const store = featherweight('prefs', {
  endpoint: 'https://YOUR-broker.run.app',
  apiKey: 'fw_...', realtime: true,
})

store.load((data, meta) => render(data))
store.save({ theme: 'dark' })

The zero-trust embedding API

Turn text into vectors for search and RAG. OpenAI-compatible, three quality tiers, owned GPU infrastructure. Sign in with Google to get an API key — the first 10M tokens are free.

curl https://api.voxell.ai/v1/embed \
  -H "Authorization: Bearer $VOXELL_KEY" \
  -d '{"texts":["how do I reset my password?"],"model":"turbo"}'

Product details on the Forge page. Plugging into n8n, Flowise, Dify and more? See the recipes.

Upload a document. Ask a question.

Voxell Answers chunks, embeds, and indexes your documents, then answers questions grounded in them with citations — in two API calls. Load, then ask. The corpus is created on first upload.

# 1. load a document (auto-chunks + embeds)
curl https://api.voxell.ai/v1/wield/handbook/documents \
  -H "Authorization: Bearer $VOXELL_KEY" \
  -d '{"name":"policy.md","text":"..."}'

# 2. ask a question, get a grounded answer + sources
curl https://api.voxell.ai/v1/wield/handbook/query \
  -H "Authorization: Bearer $VOXELL_KEY" \
  -d '{"query":"How many vacation days?","mode":"answer"}'

More on the Voxell Answers page. (Pipeline stage: Mine → Forge → Wield.)