Voice AI & Edge Hardware
Speech-driven systems that run on your own hardware
Voice is the easiest interface there is for people whose hands are busy — a warehouse, a shop counter, a workshop. I build speech systems on open-source models that run on your own machines, so there is no per-minute cloud bill and no recordings leaving your building.
What you end up with
- Hands-free capture where typing is impractical
- Speech models running on hardware you own, with no usage meter
- Audio and transcripts that never leave your network
- Low-cost devices doing real work on the shop floor
- Works on a weak internet connection, or none at all
Signs you need this
If two or three of these sound familiar, it is probably worth a conversation.
- Your staff have their hands full and typing is genuinely impractical
- Cloud speech APIs would cost more than they save at your volume
- Recordings of your operations must not leave your premises
- Connectivity at the site is unreliable but the work still has to happen
- You want smart devices doing real work rather than a novelty demo
What is actually delivered
Not a statement of intent — the concrete artefacts and outcomes you receive.
Self-hosted speech stack
Speech-to-text and text-to-speech running on open-source models — Whisper for listening, Piper for speaking, and a local language model where one is needed. Your servers, your data.
Edge devices
Low-cost single-board computers with a microphone and speaker, deployed where the work happens, talking to a central server over a documented protocol.
Real-time audio handling
Voice activity detection, streaming over websockets, wake-word handling and noise tolerance — the unglamorous engineering that decides whether a voice system is usable or infuriating.
Home and building automation
Devices, sensors and voice control tied into an automation hub, self-hosted on your own cloud VM rather than a vendor's subscription.
Deployment and monitoring
Containerised deployment, remote updates and health monitoring, so a device in another city can be looked after without a site visit.
How the engagement runs
- 01
Scope
Work out honestly whether voice is genuinely better here than a screen or a scanner.
- 02
Prototype
One device, one flow, tested in the real room with the real background noise.
- 03
Harden
Handle accents, noise, dropped connections and the ways people actually speak.
- 04
Deploy
Roll out to more devices with central monitoring and remote updates.
Voice AI & Devices: common questions
Better than most people expect, but it has to be tested in your actual room rather than trusted from published figures. Microphone choice and placement usually matter more than which model is used, which is why prototyping happens on site.
Depends on how it is built, and that is a design decision made up front. Processing can run entirely on local hardware so the system keeps working offline, or on a central server if that suits better. Either way, connectivity failure should degrade the system, not stop it.
It is practical where hands are genuinely busy and the vocabulary is bounded — a warehouse, a counter, a workshop. It is a poor fit for open-ended conversation or noisy environments with complex commands. I will tell you which situation you are in before you spend anything.
Three reasons: cost stops scaling with usage, your audio never leaves your network, and the system keeps working when the internet does not. Open-source models are now good enough that the trade-off usually favours self-hosting — though for low volumes a cloud API is simpler, and I will say so.
Less than people expect. A Raspberry Pi with a decent microphone works well as a listening endpoint, with the heavier models running on one central server or a modest cloud VM. Total hardware cost per location is genuinely small.
Reasonably well, and it is exactly the thing to test before committing. I always prototype with your actual staff in your actual environment, because published accuracy figures and real shop-floor accuracy are very different numbers.
Voice AI & Devices in your industry
The problems look different in each of these, so the way I approach them differs too.
Services that usually go with this
ERP Implementation & Consulting
From selection to go-live, without the six-figure disaster
Business Process Automation
Remove the manual work that quietly eats your margin
Accounting, Reconciliation & Revenue Assurance
Stop losing money you never knew you were losing
Want a straight answer on your situation?
Thirty minutes, no pitch. I'll tell you what I would do and what it costs.