BYOK billing: why we charge $19.99 when competitors charge $200

April 22, 2026 · Claudiu · 5 min read

Here is the dirty secret of AI orchestration pricing.

When you pay $200 per month for a multi-agent platform, you are not paying $200 for orchestration. You are paying somewhere between $10 and $40 for the actual software, and $160 to $190 to resell you OpenAI and Anthropic tokens at a markup.

The vendor buys tokens wholesale at the API provider's commercial rate. You get charged the "retail" rate baked into your subscription. The difference is their margin — and it is massive, because API tokens are where the real cost of multi-agent systems lives.

We think that model is broken. So we built a different one.

What BYOK actually means

BYOK stands for Bring Your Own Key. Instead of NEXUS PRIME buying API tokens from OpenAI / Anthropic / Google and reselling them to you, you plug in your own API keys — issued directly to you by the providers, billed directly to your card by the providers.

When one of your agents calls GPT-4, the cost shows up on your OpenAI invoice, not ours. When another agent calls Claude Opus, it goes on your Anthropic bill. When a third hits Gemini, Google charges you directly.

We never touch your token spend. We do not mark it up. We do not meter it.

Our subscription only pays for two things: the orchestration software itself, and the server fleet running the coordination layer (the council, the memory, the routing).

The math that makes $19.99 possible

Let's break down a realistic month of heavy multi-agent work. Say you run 200 orchestrated tasks, average 6 agents per task, average 15k input tokens + 3k output tokens per agent call. That's roughly 21.6M input tokens and 3.6M output tokens across the month.

On the flat-rate SaaS model, the vendor has to price at least $200 to cover those API fees plus their margin. They do not know your usage patterns in advance, so they price for the heavy user and give the light user an overpriced product. Everyone loses except the vendor.

On the NEXUS BYOK model:

Same work. Same 100+ agent capacity. Same orchestration quality. 40 to 60 percent cheaper, and your costs are fully transparent because you see them on your provider invoices.

If you run a lighter month — say 50 tasks instead of 200 — your API bill drops proportionally. On the flat-rate SaaS, you'd still pay $200. On NEXUS, you'd pay $19.99 plus maybe $30 in API usage. That is a 10x cost advantage for light months.

Why the flat-rate model exists at all

Because it's simpler to sell. "Pay us $200 and do not worry about APIs" is an easy pitch. It gives users predictability. It lets the vendor hide their cost structure. It lets them charge whatever the market will bear.

The problem is that predictability is fake. The vendor still meters you behind the scenes — rate limits, "fair use" policies, sudden overage charges when you exceed opaque thresholds. You thought you bought unlimited; you actually bought a ceiling with a hidden ladder.

BYOK flips that. You already have a relationship with the API provider. You already see the invoices. You already know what a token costs. You just plug that existing relationship into NEXUS and now you have orchestration on top of it.

What this means for what you can build

When the orchestration layer is 10 to 20 dollars instead of 200, a bunch of things change:

You can run multiple parallel workflows

A 200-dollar SaaS makes you pick one use case to justify the cost. A 20-dollar SaaS means you can spin up orchestrations for writing, research, legal review, financial analysis, customer research — all in parallel, all on the same subscription.

You can experiment without committing

Want to test whether multi-agent beats single-agent for your particular task? On flat-rate SaaS, you're locked into the full subscription to even try. On BYOK, you pay 20 bucks for a month of access and whatever your experiments actually consume in API calls. If they eat $15 of tokens, that's the total cost.

You can give orchestration to your whole team

A 20-dollar per user tool is something every knowledge worker on a team can have. A 200-dollar per user tool is an executive-only budget item. NEXUS is built to be the former, not the latter.

Your token caps are yours to set

OpenAI, Anthropic, Google all let you set monthly spend caps on your API keys. On BYOK, those caps protect you directly. You decide your ceiling. The vendor does not get to change it.

What NEXUS actually charges for

Three things, and only three things:

  1. The orchestration engine. The hierarchical planner, the PM agents, the council debate system, the scoped context routing, the quantum cloning memory layer. The stuff we wrote about in the last post — that's what $19.99 pays for.
  2. The coordination server fleet. Five owned Ollama servers handle the council debate and the free-tier model calls. These cost us about $15,000 a month regardless of how many customers we have. Subscription revenue pays that bill.
  3. Free-tier models for everyone. Even Eco-tier users get access to the owned models — no API key required for those. You only need your own keys when you want to invoke GPT-4, Claude Opus, Gemini Pro, and the other premium models.

Three tiers, three price points:

At every tier, your API spend goes to your API provider. We never touch it.

The honest pitch

Flat-rate AI SaaS will always be able to quote you a single number on a pricing page. That number will always include a hidden markup on tokens you could buy cheaper yourself.

We think the right frame is: pay us for the orchestration, pay the model provider for the models, and keep the relationship with each one transparent. You get a better product, we get aligned incentives (we only win if our software is actually valuable), and nobody is hiding anything.

That's why we charge $19.99 when competitors charge $200.

Next post: "Specialist vs generalist agents: why 100 small experts beat 1 big brain" — how narrow-scope specialist agents outperform general-purpose ones on real work, and why NEXUS is built around that bet.

Join the NEXUS PRIME waitlist

Be first in line when pre-orders open.

Claim your spot

← All posts