Our Business Model.

Small. Red. Extremely suspicious of everyone.

We have a mascot. She's small. She's red. She's extremely suspicious of everyone.

Her name is Tally.

We chose a dragon deliberately because what do dragons do? They hoard treasure. And what is the treasure of the modern world?

Tokens.

Every word you send to an AI model costs you treasure. Every bloated prompt, every wasteful tool call, every time a model three sizes too large answers a question a model three sizes smaller could have handled perfectly, that's your treasure disappearing into someone else's vault.

Tally the dragon sits on your tokens. She guards them with her life. She eyes every large language model and every MCP with deep ancestral suspicion. And she breathes fire on waste.

She's not greedy for herself.

She's greedy for you.

Show me the incentive and I'll show you
the outcome.

We read that and built our entire business model around it.

Here's the problem with almost every infrastructure company you've ever used:

Their incentive is to charge you more. Lock you in. Expand the contract. Upsell the tier. The bigger your bill, the happier they are.

Which means their happiness and your happiness are fundamentally opposed.

We thought that was insane. So we did something about it.

Opacity is how companies decay.

There was a time when Google's informal motto was "Don't be evil."

That built trust. It made people feel aligned with the company behind the search box.

Who trusts Google now?

Not because they became villains. Because the incentive shifted. Search results became monetised terrain. Every pixel competed for revenue. Every ranking decision quietly optimised for extraction.

Opacity followed. Trust eroded. The outcome matched the incentive.

We refuse to let that happen here.

We are going to publish our shapes.
We are going to publish our routing logic.
We are going to publish our judgments.

We will anchor critical telemetry and decision proofs into the blockchain so they cannot be quietly rewritten.

If we ever drift from usefulness toward extraction, the evidence will be visible.

Opacity is how companies decay.
We choose transparency instead.

But we are not naïve.

We are building this with a very specific fear in mind.

That someone bigger, louder, and better funded will try to take it from us.

History says they will.

So we designed a poison pill.

And we swallow it gleefully.

We cap ourselves. We narrow our margins. We publish our logic. We anchor our decisions publicly.

We make it structurally irrational to compete by extraction.

Because here is the trade we are offering:

Your trust in the little guy, seasoned in building systems like this…

versus your trust in the big guy who has already shown you, repeatedly, what happens when incentives drift.

We are betting that aligned incentives beat size. That transparency beats scale. That discipline beats dominance.

We don't want to be powerful.

We want to be useful. And well fed.

And we want to be structurally unappealing to competitors.

We don't take too much. We don't squeeze margins. We don't create artificial scarcity.

There's no monopoly rent to attack. No excess to undercut. No bloated margin inviting invasion.

Just a dragon doing her job.

Small. Red. Efficient. Fed.

We support telemetry wherever
the compute lives.

Our alignment goes even further.

We will support telemetry not just between the large frontier models.

We will support telemetry between your own local models.

The small model running on your laptop.
The fine-tuned model on your own server.
The air-gapped instance in your own rack.

If it is cheaper, faster, or simply smarter to run the load locally, we will tell you.

If the right answer is not to call an expensive frontier model, we will say so.

If the right answer is not to route through us at all, we will make that visible too.

Because our incentive is not to maximise model calls.

Our incentive is to maximise savings.

And sometimes the most profitable routing decision for you is the one that bypasses us entirely.

That is not a bug in our business model.

That is the proof of it.

We are not building a toll booth.

We are building an intelligence layer that optimises for your treasure, wherever the optimal compute lives.

Cloud. Edge. Local. Hybrid.

If the dragon can save you more by staying quiet, she stays quiet.

A small percentage of what
we save you. That's it.

No flat fees designed to extract maximum value regardless of whether we helped you. No enterprise tiers with features deliberately withheld to force an upgrade. No contracts designed by lawyers whose job is to make sure you can't leave.

Just a small, honest percentage of the treasure we protect for you.

The more you save, the more we make.

Which means every single thing we do — every routing improvement, every context optimisation, every wasteful token we eliminate — makes us both richer simultaneously.

Our greed and your greed are the same greed.

That has never been true of an infrastructure company before. We checked.

One cent to be listed.
Everything else is proof.

We're building a marketplace of MCPs — tools that plug into your AI workflows and make them more powerful.

We charge one cent to be listed.

Not because we don't value our marketplace. Because we want every MCP in existence to be in it. Because the more tools in the marketplace, the more calls flow through Tally, the more treasure the dragon guards, the more we all make together.

Every MCP comes with its telemetry data — real usage data showing exactly how it performs in the wild. No marketing. No promises. Just proof.

The dragon doesn't do promises. She does receipts.

The more you pay us,
the more money you make.

Here's something most companies would never say out loud.

We're saying it out loud. Because it's true. And because we think you deserve to know exactly what kind of company you're dealing with before you trust us with your infrastructure.

The math is simple:

There is no point in that loop where our interests diverge from yours.

That loop is the whole company.

We're going to be disgustingly successful.

Not because we're going to get greedy. But because we're never going to get greedy.

Scale does the work. A tiny percentage of an insanely large number is still an insanely large number. We don't need to take more. We just need the number to keep growing.

And the number keeps growing as long as the dragon keeps doing her job.

Small. Red. Suspicious of everyone.

Fiercely, completely, mathematically on your side.