Falcon Private AI
Private AI that lives on your PC.
While other traders hand their journals, statements, and unpublished strategies to cloud chatbots every day. This is the version where none of it leaves the room, there is no monthly bill, and it runs with the network cable unplugged.
You own the machine. You own the model. You own the data.
Your data stays yours
No subscription, ever
Works offline
Lifetime US support
Built in. Working the day it arrives.
All of this runs on your machine
- Review your trade journal for patterns
- Read a broker statement back in plain English
- Write and debug your indicators
- Check a trade against your own rules
- Ask it anything else, like any assistant
Nothing goes out unless you say so. Two optional features can reach the internet: a research lookup, and checking Falcon for an updated model. Both ship switched off, behind their own toggles, and neither send your files or conversations anywhere unlesss told to.
Building trading systems
Systems sold to traders
Added monthly cost, forever
Largest capacity in the lineup
Why private AI
Nobody has to hack anything for your conversations to end up somewhere you didn't choose.
This is the part most people have not thought through. A vendor getting breached is only one risk among several. Once your data lives at someone else’s address, what happens to it next is decided by people who have never heard of you: courts, regulators, their vendors, whoever owns the company in five years. None of that requires an attacker, and all of it has already happened.
The only data that can't be leaked, subpoenaed, reviewed, or resold is the data that was never given away.
Your trade journal records how you think, your statements record your entire financial position, and your unpublished strategy is the thing you actually get paid for. Exposure on any of those files cannot be undone once it happens, which is exactly why people paste them into a chat window without a second thought.
Falcon Private AI does not ask you to trust a policy. There is no account, no server but yours, and nothing to subpoena but you.
What we will not claim. Running AI yourself does not make you invulnerable, and we are not going to pretend it does. Researchers found roughly 175,000 local AI servers sitting exposed on the open internet in January 2026, almost all of them set up by someone who left a port open. A machine you own is still a machine you have to keep patched, and malware on your own PC is a risk no architecture fixes. What changes is the number of parties involved: no vendor, no vendor’s vendor, no third-party custodian, and no subpoena target other than you. Yours arrives configured, bound to your own machine, with internet access off.
Falcon Terminal, our software, pre-installed
The hardware is the easy part. This is the reason to buy it.
Anyone can put an open-source AI shell on a fast PC and figure it out over a weekend. What arrives on an AI Desk or AI Book is Falcon Terminal, software we wrote, with the work traders actually do already built in. You power it on, click a tool, and hand it a file. There is no setup, no prompt-writing to learn, and no command line to touch.
Built in, local
Review my trade journal
Hand it your fill history and it finds the patterns you cannot see in your own logs: the setups you cut short, the hours you lose money, what happens the day after a loss.
Never uploaded, never leaves the machine
Built in, local
Read my broker statement
Drop in a statement. It reads back your real concentration, your correlated positions, and what a bad week in one sector would do to the whole account.
The most sensitive file you own, no upload
Built in, local
Write my indicator
Draft and debug Pine Script, NinjaScript and thinkScript. It flags the parts it isn’t sure about rather than handing you code that looks finished and isn’t.
Your unpublished edge stays yours · AI 16GB or better
Built in, local
Check this trade against my rules
Save your own rules (risk per trade, daily drawdown cap, what has to line up before you enter) and walk a trade through them before you take it.
Your rules, saved on your machine
Built in, local
Search my own documents
Point it at a folder (prop-firm rulebooks, platform manuals, your own notes) and ask questions against it. Answers cite what they came from, and the folder never leaves your drive.
Indexed locally, nothing uploaded
Optional, ships off
Look something up online
The one tool that reaches the internet, because it has to. Switch it on and only the lookup itself goes out: a ticker, a headline search. The thinking still happens here.
Off until you enable it
And the rest of the time, it's simply your own private assistant.
Everything above is built and tuned for trading. But underneath it is a full, general-purpose AI assistant, the same kind of thing you would open a browser tab for today. Draft the email, summarize the document, explain the concept, work through the problem out loud. Same window, no account, no monthly bill, and nothing you type goes anywhere. Most owners end up using it for that as much as for the trading tools.
Falcon Terminal analyzes and explains. It does not tell you what to trade.
These tools help you understand your own trading: your history, your risk, your code, your statements. They will not produce buy or sell calls, and anything that could touch an order stays in simulation until you confirm it yourself, every time. It is the same line every brokerage draws, and it is the right one to draw.
Rent or own
Anyone can rent AI for $20 a month. That isn't what this is.
We will be straight with you: the cloud models are strong, and they are inexpensive. If all you want is a smart assistant for general questions, go subscribe to one; we do too. The reason to own a machine like this is narrower, and it only matters for a specific kind of file.
What renting means
Your account is the product.
- Your files go to their servers. Every statement, every journal, every line of strategy code you paste in.
- The default is training. OpenAI’s own help centre: “Data sharing is enabled for you by default” on Free, Plus and Pro.
- Retention outlives deletion. Google keeps human-reviewed chats up to three years even after you delete your activity.
- Courts can reach it. Deleted chats have been preserved by order, and 20 million conversations produced to opposing counsel.
- The model changes when they say. Fifteen days’ notice before GPT-4o was pulled from ChatGPT in February 2026.
What owning means
The machine is the product.
- Nothing is uploaded. The statement is read on the machine it is already sitting on. There is no cloud server, so there is nothing to send to.
- There is no vendor to train on you. No account, no history held anywhere, nobody to ask for it.
- Nothing to subpoena but you. No third-party custodian sits between you and your own files.
- You choose when the model changes. The weights are a file on your drive. They change when you replace them.
- You pay once. A ChatGPT Pro or Claude Max seat is $200 a month, which comes to $2,400 a year, every year, forever.
One fair counter, stated plainly: business and enterprise AI plans carry different terms, and the 2025 preservation order specifically excluded them. If you are on one, several of the points above do not apply to you, though on those plans your account administrator can generally read your content instead. The consumer tiers, which is what most individual traders actually use, were covered.
The lineup
Three systems, ready to roll. Pick the one that matches the capability you want.
Every one of them runs Falcon Terminal, pre-installed, with the model already loaded. The only real decision is how much AI memory you want: that is the number that sets how large and how capable a model your machine can hold.
The MI-12 is the entry point into Falcon’s I-Series: a full Falcon trading laptop with a real AI model running on the machine itself, in our 16 or 18-inch chassis built to travel. It answers with the model already loaded, Falcon Terminal already installed, and outbound access switched off, so the journal review, the chart explanation, and the document question you’d normally take to a cloud app run on hardware you own instead. Disconnect the Wi-Fi and it keeps answering, because there’s nothing about it that depends on a connection staying up.
- Lifetime U.S. Technical Support
- NVIDIA GeForce RTX 5070ti, NVIDIA GeForce RTX 5090
Fully local
From
$3,333
- Intel Core Ultra 270K Plus
- Custom F-1 Liquid Cooling System
- 32GB DDR5 (5600MT/s)
- 1300 watt
Fully local
From
$7,343
Local, Private AI Hosted entirely by you.
Empower your workflows with absolute privacy and raw local compute. Packed with 128GB of high-speed LPDDR5X unified memory, the Falcon Enclave AI is built to handle complex, multi-agent orchestrations without ever sending a single byte to the cloud.
Key Use Case:
-
Strategic Orchestration: Seamlessly load massive foundational models to draft high-level blueprints and break down complex problems.
-
Dynamic Agent Swarming: Offload and swap into multiple specialized, smaller AI agents on the fly to execute, calculate, and solve sub-tasks in parallel.
Your data stays on your hardware. Sovereign AI performance, right on your desk.
- AMD Ryzen™ AI Max 300 Series
- 128GB LPDDR5x
- 2TB NVMe Gen 4 SSD
Fully local
From
$6,390
AI Desk Duo and AI Book Duo
The best setups come as a Duo.
If you’re buying one of these, this is usually the answer. A Duo pairs an AI Desk or AI Book with a Falcon AI Enclave, sold together at a discount, and you get both halves of what local AI should feel like.
Your Desk or Book handles the everyday work: fast, responsive, right there in front of you. The Enclave sits alongside it holding the largest models for the heavy jobs: a full earnings transcript, a year of statements, the long reasoning you do not want to wait on a smaller model for. One toggle in Falcon Terminal picks which one answers.
Your network, stays here
everyday, instant
the big models
Away from your desk. The Enclave stays home, so off your own network you fall back to the model running on the laptop itself, still enough for the journal, the statements, and most day-to-day work. We would rather say that plainly than sell you a tunnel back to your house.
No rebuild
Your current machine stays exactly as it is
Buy once
One Enclave serves every Falcon you own
Shared
Two or three at a desk can work against one
Capability
What each build will actually do for you
AI memory is the number that matters: it sets how large a model your machine can hold, and model size decides how much it knows and how carefully it reasons. Here is what each step buys you, in the work rather than the spec.
Provisional: being re-measured on the shipping software
Build
AI memory
Model class
What that means in practice
AI Book (entry)
12 GB
8–12B
Quick, fluent everyday assistant. Reviews your journal, explains code and charts, handles shorter documents. Not the one to buy if you want it writing indicators from scratch.
AI Desk (base)
16 GB
14B dense · 20B sparse
The honest floor for AI that writes strategy code. Comfortable with full statements and long documents, and quick enough to keep up with you all day.
AI Book (18″)
24 GB
up to 24B
Everything above, in a laptop. Larger documents and more careful reasoning than the entry Book, with no connection needed.
AI Desk (mid)
32 GB
32B dense · 30B sparse
A noticeably sharper model. Better at multi-step reasoning, longer strategy code, and holding a big document in mind while you ask about it.
AI Desk (ceiling)
64 GB
Being validated (provisional)
Headroom for the largest models a graphics card can hold, and for keeping more than one model loaded at a time.
Falcon AI Enclave
128 GB
120B-class sparse (5.1B active)
The strongest reasoning we offer. Takes a full earnings transcript or a year of statements in one pass, and handles questions a smaller model would answer shallowly.
Either Duo
128 GB + your machine
both, on demand
Fast local model for everything routine, the Enclave’s big model when the job deserves it. One toggle decides which answers.
What "14B" or "120B" actually means.
8–14B
Fast and genuinely useful. Summarizing, explaining, reviewing your journal, answering questions about a document you hand it. Occasionally shallow on hard multi-step problems.
14–32B
Where most traders should land. Writes usable strategy code, reasons through a statement properly, and keeps its footing on longer documents.
120B-class
The Enclave’s territory. Full transcripts, dense filings, questions with several steps in them. The closest thing to a frontier model that will run on your own desk.
It is the size of the model, counted in billions of parameters, roughly how much it knows and how carefully it thinks. Bigger models reason better and hold more of a long document in mind at once, and they need more memory to do it. That is the whole trade, and it is why AI memory is the only spec on this page we ask you to care about.
Want the full explanation: what AI memory really is, what speed means, whether local AI is genuinely free and actually private? Read the buyer’s guide to local AI →
Go deeper
Everything else worth reading before you decide
Buyer’s guide
Local AI, explained properly
What AI memory is. What parameters mean. What speed actually measures. Whether it’s really free, and whether it’s actually private. Start here if any of this is new.
Software
Falcon Terminal
The application that ships on every machine: every built-in tool, how the privacy toggle works, and what it does when you are not trading.
Hardware
Configure a machine
AI Desk, AI Book, and the Enclave in full: every configuration, what each upgrade actually changes, and which one we would put you on.
Straight answers
Before you configure anything
By default, no. Falcon Terminal runs entirely on your machine and needs no connection to answer a question. You can unplug the network cable and it keeps working.
Two features can reach out, and both ship switched off: pulling a model update, and a research lookup. They are one toggle inside Falcon Terminal, named on this page and in the app rather than buried in a settings menu, and neither sends your files or chat history out. If you would rather it were enforced in hardware than trusted to software, ask us to block the AI’s network access at the firewall.
We would rather give you the accurate answer than the scary one. The major US AI vendors have a good security record: there has been no mass exposure of conversation content at OpenAI, Anthropic, Google or Microsoft. Incidents have involved account metadata, or vulnerabilities that researchers found and vendors patched before anyone exploited them.
This is exactly why we do not build the argument on breaches. The documented risk takes other forms: a court order, a vendor’s vendor, a retention policy that outlives your delete button, a company whose ownership and terms can change. Those have all already happened. A breach would only be one more route to the same place.
For a lot of what you would ask an AI, you should. It is inexpensive and it works well, and we are not going to argue otherwise.
The question is what you are comfortable sending. Explaining a butterfly spread: send it. Uploading your account statement so it can tell you where you are over-concentrated, or your unpublished strategy so it can debug it, is the one that gives people pause. This machine is for that second category, and for the fact that you own it outright instead of renting access to someone else’s.
No, and we are not going to pretend otherwise. The frontier cloud models are ahead of anything that runs on a desk, and they will stay ahead.
What runs on your Falcon is capable enough for the work that matters: reviewing a year of trades, explaining a spread, drafting an indicator, or reading a statement back to you in plain English. The case for running it here is what you can safely hand it, your account statements, without thinking twice about where they end up. Most of our customers end up using both, and staying deliberate about which questions go where.
On raw infrastructure, often yes, and we will not tell you otherwise. Their buildings have security teams; ours is a computer in your office.
But that is not the comparison that matters. A datacenter protects the data it holds; it does not stop the data being held. What owning changes is how many parties are involved: no vendor, no analytics supplier, no third-party custodian, and nobody to serve a subpoena on except you. Nearly half of all breaches now arrive through somebody’s third party, an entire category of risk that simply does not exist when the file never leaves your desk.
No. It arrives with Falcon Terminal installed, tuned to your specific machine, and loaded with the models we have selected for it. You open it, click the tool you want, and hand it a file. There is no command line, no account to create, and nothing to configure.
On an AI Desk, no, and not as a matter of tuning. Your monitors are driven by graphics hardware that does nothing else; the AI runs on separate hardware that is never asked to draw anything. There is nothing shared to contend over.
An AI Book has one graphics chip doing both jobs. That is fine for a question here and there, and it is exactly why the AI Desk exists, or why pairing a Book with an Enclave removes the problem entirely.
You load it, if you want it. New models come out constantly and they run on the same hardware; we evaluate the ones worth caring about against the work traders actually do, and release the ones that earn it for your machine.
Nothing installs itself and nothing changes underneath you, which is the opposite of the cloud problem, where the vendor can retire the model you had gotten used to on a couple of weeks’ notice, or apply an update that changes the output, overnight.
Not sure which one you need?
Tell us what you want the AI to do (review your journal, write your indicators, read your statements) and we will tell you the smallest machine that does it well. That is a real conversation with a technician, not a form.