Private AI.
By application only.

Real open models — up to 33 billion parameters — running on hardware we own in Cleveland, Ohio. Not a cloud. A room. And the room is small.

Membership is limited · admission by application

Talk to the stand-in

This demo is a scripted simulation — honest about it, too. The real models come with a seat.

Demo · scripted simulation · not a live model

The models

Weighted by size: light models sip your monthly tokens, heavyweights drink them. Every seat includes the Core library; bigger tiers unlock bigger minds — or buy a single model à la carte.

Llama 3.2 (3B)
Our lightest model — the free taste.
Free taste · 0.5× tokens
Llama 3.1 (8B)
Fast all-rounder.
core · 1× tokens
Mistral (7B)
Quick and direct.
core · 1× tokens
DeepSeek R1 (8B)
Compact reasoner.
core · 1× tokens
Qwen 3 (14B)
Sharp general model.
plus · 2× tokens
DeepSeek R1 (14B)
Step-by-step reasoning.
plus · 2× tokens
Phi-4 (14B)
Microsoft's dense thinker.
plus · 2× tokens
Qwen Coder (14B)
Code-focused.
plus · 2× tokens
Gemma 4 (9B)
Google's open model.
plus · 2× tokens
GPT-OSS (20B)
Big, deliberate.
max · 4× tokens
Gemma 3 (27B)
Deep generalist.
max · 4× tokens
Qwen 3 (30B)
Heavyweight reasoning.
max · 4× tokens
DeepSeek Coder (33B)
Our largest code model.
max · 4× tokens

Pricing

Monthly. Cancel anytime. Every membership holds one of a strictly limited number of seats.

Core
$15/mo
  • 1.0M weighted tokens / month
  • The Core library (7–9B models)
  • Fast responses, private by design
  • Token top-ups & single-model unlocks
Apply — $5
Plus
$25/mo
  • 2.5M weighted tokens / month
  • Everything in Core
  • The 14B library
  • Token top-ups & single-model unlocks
Apply — $5
Max
$40/mo
  • 5.0M weighted tokens / month
  • Everything in Plus
  • The deep library (20–33B)
  • Token top-ups & single-model unlocks
Apply — $5

Admission is by application: a one-time $5 fee, credited toward your first month if you're accepted. Weighted tokens: a reply's tokens × the model's multiplier. USD via PayPal.

Questions people actually ask

This list rearranges itself: the demo bot keeps count of what visitors ask, the popular questions rise, and the ignored ones quietly leave.

What does membership cost?

Core is $15/month, Plus is $25/month, Max is $40/month. The difference is which models you unlock and how many weighted tokens you get. Cancel anytime.

How do I apply?

Fill out the short application on this page and pay the $5 application fee. Applications are reviewed personally — you'll hear back by email.

What is WeldonPC.ai?

A private AI membership. Real open models — up to 33 billion parameters — running on hardware WeldonPC owns in Cleveland, Ohio. Membership is limited and admission is by application.

Is the demo a real AI?

No — this demo is a scripted simulation, and proudly says so. Members talk to the real models. The honesty is the point: you always know which one you've got.

Why is there a $5 application fee?

It keeps applications serious — membership is limited, and review time is real. If you're accepted, the $5 is credited toward your first month.

Is the application fee refundable?

If you're accepted, it's credited toward your first month. If you're declined, the fee covers the review and isn't refunded.

What happens after I apply?

Your application is reviewed personally — no algorithms deciding your fate here. You'll get an email either way, usually within a few days.

Why is membership limited?

Because the hardware is real. Every member shares one privately owned GPU server, and we keep the room small enough that it stays fast for everyone in it.

What's the difference between tiers?

Core ($15) unlocks the fast 7–9B library with 1M weighted tokens. Plus ($25) adds the 14B library and 2.5M tokens. Max ($40) adds the deep 20–33B library and 5M tokens.

How do weighted tokens work?

Every reply's tokens are multiplied by the model's weight — light models are 0.5–1×, 14B models 2×, the deep library 4×. Small questions to small models barely dent your month; big minds drink deeper.

What happens when I hit my monthly limit?

Chat pauses until your cycle resets — or you grab a token top-up ($5 per 500k weighted tokens) and keep going. No surprise overage charges, ever.

Can I buy more tokens?

Yes — $5 adds 500k weighted tokens whenever you need them. They stack on top of your monthly allowance.

Apply for membership

A short application and a one-time $5 fee — credited toward your first month if you're accepted. Every application is reviewed personally.

Where your words go

The short version, without the lawyer smell.

They stay on our hardware. Your conversations run on a machine WeldonPC owns in Cleveland. No OpenAI, no Google, no third-party cloud in the path.
They are never sold, and never train a model. Not ours, not anyone's. The models are open weights, running as they came.
Staff can read them when there's a reason. Support, billing disputes, abuse, or a lawful request — and every time a conversation is opened, it's recorded. A small shop you can call, not a black box.