PHNTM
Local AI and cloud services

What leaves your house.

The full local-vs-cloud boundary is on the privacy page. Head-to-heads: vs Alexa · vs ChatGPT.

PHNTM ONE · ALEXA ON ECHO · CHATGPT

Choose the tool for the job. PHNTM One is dedicated local AI hardware. Alexa on Echo is a voice service delivered through a speaker. ChatGPT is a cloud-hosted chatbot. These serve different needs; the table compares their relevant settings and dependencies, not identical products. For DIY local AI, see the buyer’s guide.

PHNTM prioritizes local use and inspectable data. Alexa and ChatGPT have their own privacy controls and strengths. Read both the advantages and the trade-offs before choosing.

WORKS WITH THE INTERNET UNPLUGGED
phntm one
yes — local chat, memory, voice and documents. Reminder timing depends on a verified clock
Alexa / Echo
Voice requests use Amazon’s cloud; other Echo functions vary by model
ChatGPT
no
WHAT LEAVES YOUR HOUSE
phntm one
Private-mode local chat does not send content to a cloud model. Connected accounts have separate paths. The Device screen shows observed connections, though attribution can be incomplete and short connections missed; see the limits
Alexa / Echo
Echo requests activated by a wake word or action button go to Amazon for processing under its published Alexa terms
ChatGPT
Chat requests are processed by OpenAI; training depends on your settings
THE DAY THE COMPANY SHUTS DOWN
phntm one
core local functions do not depend on DeCampLabs servers (the continuity page)
Alexa / Echo
Cloud features depend on the provider continuing its service
ChatGPT
Cloud-hosted answers depend on the provider continuing its service
CAN YOU READ THE CODE IT RUNS?
phntm one
yes — readable Python, on hardware you own, SSH documented
Alexa / Echo
no
ChatGPT
no
REMOVE AN ACTIVE MEMORY
phntm one
Library → Memories stops active memory recall; index cleanup may finish later if unavailable. Conversations, documents and older copies are separate. Read deletion limits
Alexa / Echo
Deletion is managed through the provider’s privacy controls
ChatGPT
ChatGPT offers memory and chat deletion controls; retained copies follow its published policies
THE MICROPHONE, WHEN YOU'RE NOT TALKING
phntm one
tap-to-talk by default; local wake detection is optional and off until enabled. Tap to talk: the red dot shows while it records, and it stops after about 5 seconds of quiet, a second tap or after 45 seconds; Notes dictation records up to 5 minutes with an on-screen timer
Alexa / Echo
Echo detects its wake word locally when its microphone is enabled, then sends the request to Amazon
ChatGPT
n/a
COST AND SERVICES
phntm one
$799 one-time · plus applicable tax. No required subscription for local use.
Alexa / Echo
Lower hardware cost; optional subscriptions and services vary by product
ChatGPT
ChatGPT has free and paid plans; compare the plan you would actually use

Where they win — honestly

THE ROWS WE LOSE
raw intelligence
Frontier cloud models are stronger at complex reasoning than the small local model; PHNTM labels which brain answered
speed
Cloud response speed varies. PHNTM One’s local answers can take tens of seconds; a long answer or a model reload can take more than a minute, and the first answer after switching on can take up to about 2 minutes. On November 17, 2026, three local replies in an ongoing review-unit chat after a meeting showed their first sentence in 23–30 s and finished in 30–37 s. A model reload (after switching on, after a repair or when memory runs short) can be slower. These measurements do not predict your unit’s response time; see the test conditions.
integrations
Alexa offers smart-home and service integrations. PHNTM offers optional cloud and account connections
sticker price
An Echo speaker has a lower hardware price; its purpose and services differ from this dedicated local AI device

Choose the tool that fits how you work. PHNTM One brings local AI, editable memory, your documents and voice together in a dedicated computer you own. Cloud services have different strengths, especially speed and complex reasoning.

Is this your kind of AI?

Good fit if you value local control, a dedicated desk device, and room to develop notes, memories and file questions. Less suited if you need instant answers, frontier-level reasoning, guaranteed factual accuracy or a battery-powered tablet: PHNTM One has no battery and uses an included power supply. Its local model can take tens of seconds; a long answer or a model reload can take more than a minute, and the first answer after switching on up to about 2 minutes. Compare the trade-offs and read the measured conditions before ordering.

Ready when you are

Own your AI. Keep your words.

Order your PHNTM One. Each unit must complete an automated burn-in on its shipping build before shipment. $799 once, with no required subscription for local use.

$799 one-time · ships within 6–8 weeks after payment (EU: pre-order) · free US shipping · ships worldwide · 30-day returns

Comparison sources checked October 21, 2026: Amazon on local wake detection, Alexa terms, and OpenAI on data controls and Temporary Chat. ChatGPT allows training opt-out and Temporary Chat; cloud processing and training are different questions. Specific plans, settings and devices matter.