Prometheus reaching for the Tokkey cube
TokkeyLocal AI, in your hand
The Prometheus Edition

Fire, brought down to your desk.

The gods kept intelligence in the cloud. Tokkey carries it home: a small cube that runs your AI on your own machine, with your data staying where it belongs.

Your model. Your keys. Your fire.

Move to tilt. Hover the light to pull the flame.Scroll
What's new this edition

Three gifts from the fire.

Everything the cloud kept for itself, brought down to a cube that plugs into the machine you already own. Nothing here asks you to rent back what was yours.

01 / Ownership

Intelligence you keep

Models run on your hardware. Prompts and files never leave the machine unless you send them.

02 / Spark

Plug in. Light up.

Tokkey pairs with your Mac in seconds and sets up the runtime for you. No terminal needed.

03 / Flow

One key, every app

A single local endpoint your tools already speak, so the fire reaches every workflow.

How the fire travels

From the cloud, down to the desk.

Each move is a decision the cloud used to make for you, handed back. The interfaces below are working previews.

Assistanton-device via Tokkey

Tokkey
TokkeyConnected ↗
Local modelQwen 3.5 9B ↗
RouterOn ↗
Own your intelligence

Turn your Mac into an AI Mac

Plug in the cube and an assistant answers on your own machine. The menu bar shows what is running, which model, and whether anything left.

See it working ↗
Models on this Mac
Qwen 3.5 9BLoading into memory…
On disk
Stored on your SSD. No session expires it.

The weights come home

A model that lived in a data center now lives on the disk you own. It loads once and stays.

How loading works ↗
Setting up
  1. Tokkey plugged in
  2. Hardware detected24 GB unified memory
  3. Model picked to fit
  4. Endpoint readylocalhost:8080

The runtime lights itself

No terminal, no account. The runtime finds the hardware, picks a model that fits, and starts a local endpoint.

Setup guide ↗
Terminal
EditorNotesChat

Your tools already speak it

One local endpoint that your editor, your notes, and your terminal can call. The fire reaches the workflow you already have.

Connect an app ↗
Privacy
Sent off this device today0 requests

Nothing leaves unless you send it

Prompts, files, and memory stay on the machine. When a task needs more, the route out is visible, named, and reversible.

Privacy controls ↗
Where the model runs
Who handles the prompt
Cost over time
When the provider changes
Getting started

The same task, two places to run it

Switch between the two columns to see what changes when the model lives with you.

Compare in detail ↗
TokkeyThe fulcrum, on your desk
The Archimedes Edition

A lever, long enough for the world.

Archimedes asked for one firm place to stand. Tokkey is that place: a fulcrum on your desk that moves models too large to hold, without ever handing them to the cloud.

Give me a point to stand, and I will move it.

Move to tilt the scene.Scroll
What we work on

The research, drawn out.

OneToken builds the infrastructure that brings intelligence home. The sketch underneath is the idea. The work on top is the research.

Inference01

Serving that fits the machine

Scheduling, batching, and KV cache management for models running on the hardware people already own, where every megabyte of memory is spoken for.

Storage02

Weights that move fast

SSD-aware loading and caching, so a model that does not fit in RAM can still start in seconds instead of minutes.

Training03

Learning, on device

Fine-tuning and distillation small enough to run where the data already lives, instead of shipping it to a cluster.

Infrastructure04

One runtime, every app

A local endpoint the tools people already use can speak, so the model is infrastructure rather than another window.

Secondsto a running model
On devicewhere the data already is
One endpointevery tool already speaks
The journal

What we measure, and what we publish.

Engineering notes from OneToken. Every claim has a workload and a reproducible setup behind it.

Featured · PerspectiveThe fulcrum is the point8 min read

The fulcrum is the point

Archimedes never said the lever was the hard part. The hard part was a place to stand. For local AI, the model is the lever and the machine you own is the ground.

Read the essay ↗
Scorecard same task, one run
Quality
First response
Completion
Memory
Illustrative values
Engineering · 2026-10 · 5 min

Measure what actually gets done

A faster token only counts when the system finishes the right task.

Read ↗
Intelligence density
0useful work per GB · per watt
Engineering · 2026-09 · 7 min

More intelligence, a smaller footprint

How much useful work a model does per megabyte of memory and per watt, on hardware people already own.

Read ↗
Loading Qwen 3.5 9B
SSD
RAM
Layers resident0 / 36
Ready to answer…
Systems · 2026-09 · 6 min

Weights that move fast

SSD offloading for models that do not fit in RAM, so a model starts in seconds instead of minutes.

Read ↗
Where should this run?
Summarize my meeting notes
On deviceSelected cloud model
Private file. Kept local.
Product · 2026-08 · 6 min

Local and cloud, by choice

Which tasks stay on device, which may reach a selected cloud model, and how that choice stays visible.

Read ↗
Moving to a new Mac
Nothing travels unless it is ticked.
Product · 2026-08 · 6 min

A memory that belongs to you

What should travel when an agent moves between devices, and what should not travel automatically.

Read ↗
This month
Cloud subscription$20.00
API usage$0.00
Running total$20.00
Illustrative. A bill that grows with every token.
Perspective · 2026-07 · 4 min

Stop renting intelligence

Personal intelligence should not become another monthly bill.

Read ↗
Give me a place to stand, and I will move the earth.Archimedes, as told by Pappus of Alexandria
TokkeyDrawn for one hand
The Arthur Edition

The sword that answers to one hand.

The cloud offers power to anyone who pays the rent. A sword in the stone answers only to the one it was drawn for. Your model should do the same.

Sovereignty, kept on the device.

Move to tilt the scene.Scroll
The vision

A future of your own.

Five principles we build against. Intelligence should follow the work, stay on the device you own, and never cross a boundary you did not choose.

01

Flow

流动

Intelligence follows the work. A useful system carries the right context to the right capability, and the path stays understandable to the person using it.

02

Ownership

拥有

A foundation of your own. Keep useful capability on your device. Change a model without starting your relationship with intelligence from zero.

03

Possibility

衍生

One foundation, many ways to work. Inference becomes valuable when it supports an application, a decision, or a finished piece of work.

04

Together

共生

Local and cloud, by choice. Let routine and private work stay close. Ask for additional capability only within the boundaries you choose.

05

Privacy

隐私

A boundary you can understand. Inspect the source, select what may be shared, and keep the ability to stop. Privacy is a set of controls, not a promise.

What sovereignty looks like

A model that answers to you.

Ownership is a set of things you can see, change, and take with you. Try each control.

Model
MemoryKept · 1,284 notes
PreferencesKept
Switched without starting over.
The model

Yours to keep

Swap the weights without resetting your relationship with intelligence. Memory and preferences stay when the model changes.

Learn more ↗
Route for this task
ModelQwen 3.5 9B · on device
Recorded in logYes
The route

Yours to choose

Local by default. A task reaches the cloud only when you say so, to a model you named.

Learn more ↗
May leave this device
The boundary

Yours to draw

Inspect what may leave, select it, and keep the ability to stop. Privacy as controls you operate.

Learn more ↗
What Tokkey remembers about you
  • Prefers short summaries with action items
  • Works in Pacific time
  • Current project: Winter '26 launch
Visible, editable, and deletable.
The memory

Yours to correct

What the system remembers is visible and deletable. Continuity across devices never becomes hidden authority over your data.

Learn more ↗
Whoso pulleth out this sword of this stone and anvil, is rightwise king born of all England.Inscription on the stone, Le Morte d'Arthur

Questions we hear

Does Tokkey ever send my data to the cloud?

Not by default. Everything runs on the device. If you choose to route a task to a cloud model, you name the model, you see what is sent, and the choice is recorded where you can read it back.

What happens if I switch models?

Your memory, preferences, and history stay with you. The model is a part you can replace; your relationship with it does not reset.

Do I need to know how to set up a model?

No. Plug the cube in and the runtime picks a model that fits your hardware and starts a local endpoint your tools already know how to use.

Can I stop it?

Always. Turn off the cloud route and nothing is sent anywhere. Unplug the cube and the local runtime stops.

Take the fire.

Tokkey Editions. Winter '26.

Reserve your spot