The cost question, answered straight

What Bring-Your-Own-Key AI Really Costs

The short answer. Fred is $99 once for the software; the intelligence comes from an AI key you own, billed directly by your provider at their rates, with no markup from us. Light use often lands at a few dollars a month. Heavy automation can cost more than a subscription would. Anyone who tells you one flat number for "what AI costs" is selling something — so instead, this page explains what actually drives the bill, and how Fred keeps the real number in front of you at all times.

The model

How BYOK works

You create a key. An account with Anthropic (Claude), OpenAI, or Google Gemini — a few minutes, and the wizard walks you through it.
You paste it into Fred once. It's stored on your machine, never on ours. Every AI request travels directly from your device to your provider.
Your provider bills you for usage. Their rates, your account, no middleman. Fred's dashboard shows the running total in real time.

Fred works with Anthropic, OpenAI, Google, AWS Bedrock, DeepSeek, Groq, and local models via Ollama — and you can change providers or models from the dashboard whenever you like. The assistant survives any pricing decision any provider ever makes.

What drives the bill

Three dials set your monthly cost

How much you delegate. A daily briefing and some drafting is light. Long research reports, heavy coding sessions, and image generation are the expensive end — real work costs real thinking.
Which model does the thinking. Every provider sells tiers: fast-and-cheap models for routine work, frontier models for hard work. Fred lets you pick per your taste, and switching is a dashboard control, not a migration.
Cloud or local. A local model via Ollama has no per-call cost at all. The honest trade: quality then depends on your hardware, and small local models are noticeably behind the frontier. Many owners run cheap-local for routine tasks and a cloud key for the hard ones.

Honest arithmetic

When BYOK wins — and when it doesn't

Light users usually win big. If you'd use an assistant a few times a day, pay-per-use often lands at a few dollars a month — against $20 every month for a chat subscription, on top of which the subscription still can't touch your files.
Heavy automation can flip it. Run large research and coding jobs all day and your provider bill can pass what a subscription costs. You'll see it happening on the dashboard — and you can throttle, switch to a cheaper model, or push routine work to a local model.
The $99 is not the cost of the AI. It's the cost of the assistant around the AI: the dashboard, sixteen configured specialists, the phone app, workflows, permissions, support. The thinking is metered by whoever you choose to do it.

Fred dashboard with real-time cost tracking — your actual AI spend on your own key, updating live — The bill, on the glass, as it forms — not a surprise at month end.

Why we built it this way

The case for owning the meter

Bundled "AI included" pricing has to charge you for the average heavy user and add margin on top. BYOK removes the bundle: you pay your provider exactly what your usage costs, you choose the model doing the thinking, and your provider account — with its history and credits — remains yours whether or not you ever run Fred again. It's the same philosophy as the rest of Fred: one-time purchase, your hardware, your data, your key.

Pricing

The software, once

$99one-time

Every agent, every workflow, the dashboard, the Android app, and the owner's guide — from day one. No subscription. Bring your own AI key. 14-day refund window. One machine at a time.

Get Fred — $99

One-time purchase · licence key by email in minutes · 14-day refund window

FAQ

Questions people ask

What does BYOK mean?

Bring your own key. You create an API key with an AI provider — Anthropic (Claude), OpenAI, or Google Gemini — paste it into Fred once, and the provider bills you directly for what you use. Fred adds no markup and never sees your traffic.

So what will I actually pay per month?

It depends on use, honestly. Light use — questions, drafting, a daily briefing — often lands at a few dollars a month. Heavy automation — big research jobs, lots of code, image generation — costs more and can exceed a $20 subscription. Fred's dashboard shows your spend in real time, so you're never guessing.

Which providers does Fred support?

Anthropic, OpenAI, Google, AWS Bedrock, DeepSeek, Groq, and local models via Ollama. You can switch providers or models from the dashboard at any time without losing your assistant.

Can I avoid per-use costs entirely?

With a local model via Ollama there's no per-call bill — the trade-off is that quality depends on your hardware, and smaller local models are noticeably less capable than the frontier cloud models. An internet connection is still required either way.

Why not just bundle the AI into a subscription?

Because bundling means markup and lock-in. BYOK keeps three things in your hands: the real price (provider rates, no middleman), the choice of model doing the thinking, and the relationship — if you ever stop using Fred, your key and your provider account are still yours.

Before you go

Not ready to spend $99? Take the honest guide instead.

The Straight-Answers Guide to Owning Your Own AI — every hard question on these pages, answered without the sales voice. Free, by email.

One email with the guide; occasional honest notes after. Unsubscribe anytime. No spam, ever.

Keep reading