cd /news/ai-safety/hotword-detection-in-fhe-on-gpu-in-3… · home › topics › ai-safety › article
[ARTICLE · art-150697] src=jeremykun.com ↗ pub= topic=ai-safety verified=true sentiment=↑ positive

Hotword detection in FHE on GPU in 350ms

Belfort Labs launched a live demo at hotword.belfortlabs.com running a fully homomorphic encryption (FHE) hotword detector on its proprietary GPU platform, with server-side evaluation completing in 350ms, according to HEIR compiler creator Jeremy Kun. The round trip breaks down as 15ms for client-side encryption, 350ms for server-side evaluation, and 600ms for network latency, totaling under one second, though one-time setup encryption keys still take about a minute to generate in the browser and upload. The demo, compiled with the HEIR homomorphic encryption compiler, encrypts spoken audio client-side and returns an encrypted result that the client decrypts to update a word-connection game, and Belfort has opened an early access program for its FHE-based cloud platform.

read2 min views15 publishedOct 15, 2026

A quick update on our homomorphic encryption compiler HEIR: our partners at Belfort have a hotword detector in FHE that runs in 350ms.

More specifically, they launched a live demo at hotword.belfortlabs.com, that has a little game inspired by the NY Times Connections puzzle. Instead of typing in your guess for the word that connects a group of clues, you say it out loud.

Your audio is encrypted under FHE, transitted to their server, and a hotword detector runs on it to determine which word you said (still encrypted). The client receives the encrypted result, decrypts it, and updates the game state.

Not only was this demo compiled with HEIR to target Belfort’s (proprietary) GPU platform, but the inference step runs in 350ms.

More specfically, the breakdown of runtime is: 15ms for client-side encryption, 350ms for server-side evaluation, 600ms for network latency (my Portland-based laptop to Belfort’s servers which are likely in Belgium) for a total round-trip of less than one second.

The one caveat is that they pre-load the one-time setup encryption keys, which currently take a minute or so to generate in the browser and upload to the server. So if you want to play the game securely, you need that extra step.

Of course, this could all be done locally, but it’s an impressive showcase of how fast FHE has become. And to my delight, how useful our compiler has become for compiling FHE workloads!

There’s still a lot more work to do, but this milestone makes me proud of the community of academics and practitioners we’ve grown around this compiler. Hooray for collaboration!

Belfort also has an early access program for their FHE-based cloud platform which you can sign up for here. You can forgive me for the advertisement, but they intend to use HEIR as a primary tool. So if you use their platform, you’ll be helping find bugs and improvemenets in the HEIR compiler!

Want to respond? [Send me an email](mailto:mathintersectprogramming@gmail.com),
[post a webmention](https://webmention.io/www.jeremykun.com/webmention),
or find me [elsewhere on the internet](https://www.jeremykun.com/about/).

This article is syndicated on:

── more in #ai-safety 4 stories · sorted by recency
── more on @heir 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/hotword-detection-in…] indexed:0 read:2min 2026-10-15 · —