yogi.compute

Hold the token. Run every model. This document explains the mechanism, the access model, and the economics behind it.

Introduction

Every response from an AI model requires compute. Every agent consumes inference. Every product built around a model eventually answers the same economic question: who pays for the intelligence?

$YOGI is an experiment in an inference capital market: activity around $YOGI generates the capital used to continuously purchase compute for the people who hold it.

• Every $YOGI trade generates creator rewards
• Rewards are collected into a treasury
• The treasury funds AI inference for holders
• Hold 100,000 $YOGI → inference becomes available to you

The trading activity of $YOGI is directly connected to one of the largest operating costs of an AI product: inference. That relationship is the foundation of the experiment.

Quickstart

• 1. Connect a wallet — top-right corner; an embedded wallet on Robinhood Chain is created automatically
• 2. Acquire $YOGI — via the Pons Family launchpad
• 3. Hold 100,000 $YOGI — access turns on automatically once the balance crosses the threshold
• 4. Run models — from the web chat or the OpenAI-compatible API

How it works

The loop is closed by the asset itself: holding grants access, and trading activity funds the service being accessed.

trading activity
      │
      ▼
creator rewards ──► treasury ──► upstream inference bill
                        │                │
                        └──── runway ◄───┘
                              │
                              ▼
              holders ≥ 100,000 YOGI get access

Inference capital markets

For most of software history, computation was cheap enough that its marginal cost rarely mattered to the end user. AI changes that: models consume resources every time they run, and persistent agents, vision, and long-context reasoning produce enormous inference demand.

Access to intelligence starts looking less like a feature and more like an economic resource — one that needs a budget behind it. A market can generate capital, that capital can purchase compute, and that compute can power software.

Holding as access

The access model is intentionally simple: hold 100,000 $YOGI and you qualify. Your tokens stay in your wallet — there is nothing to stake or lock.

• The wallet itself is the account; the token balance is the access status
• A wallet with 100,000 gets the same access as a wallet with 1,000,000 — it is a step function, not pro-rata
• Fall below the threshold → access stops; cross it again → access returns
• API keys identify the wallet but do not permanently grant permission — the balance decides

Under pro-rata credits, access is a diagonal: near zero for almost everyone. Under a threshold it is a step — identical for everyone above it.

Treasury & rewards

Every trade generates creator rewards. A keeper checks for them approximately once a minute and claims them into the treasury, where they become capital available to finance inference.

• More rewards generated → more capital available for inference
• More inference consumed → more cost absorbed by the treasury
• Market activity finances compute while ownership determines who consumes it

API keys

Once your wallet holds 100,000 $YOGI, every model in the catalogue is available over an OpenAI-compatible endpoint. Keys are prefixed sk-yogi- and identify the wallet behind each request.

export YOGI_API_KEY=sk-yogi-...

curl https://api.yogi.compute/v1/chat/completions \
  -H "Authorization: Bearer $YOGI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-opus-5",
    "messages": [{"role": "user", "content": "gm"}]
  }'

Rate limits

Unmetered does not mean infinite. Limits are layered by what a request costs, so scarcity is financed differently — not denied.

Model catalogue

yogi.compute exposes a live catalogue — text, vision, image and video generation — over one interface. The catalogue evolves as upstream availability changes; the capital mechanism underneath does not need to predict which model wins.

• OpenAI-compatible chat completions
• Text, vision, image and video models
• Prices shown are what models cost the treasury, not what they cost you

Real-time metrics

A model like this should have to prove itself with numbers. Creator rewards, inference usage, inference cost, and treasury capital are all measurable — and from them the system calculates runway based on actual recent spending.

A market claiming to finance compute should have to demonstrate that it can actually finance compute.

Fee flow

Instead of each holder maintaining an individual credit balance, the treasury pays the upstream inference bill while creator rewards replenish the capital behind it. The user experiences access rather than a meter counting down toward zero.

Limitations

• Fees arriving — measured automatically
• Usage — measured automatically
• Runway — calculated automatically
• Conversion and top-up — performed by a person
• The treasury is an ordinary wallet on Robinhood Chain, controlled by a key

If markets are going to finance software, the financial boundaries should be visible rather than buried beneath language about automation.

Token info

FAQ

Is there a per-request quota?

No. Not a quota and not a share — every holder above the threshold gets the same access. Only the rate-limit layers above apply.

What happens if I sell?

Access is checked against your live balance. Drop below 100,000 $YOGI and requests stop being served; cross it again and access returns.

Does holding more give me more?

No. The threshold is a step function: identical access for everyone above it.

Is this free AI?

There is no such thing. GPUs cost money and providers charge for inference — $YOGI simply changes where that expense lands: on the treasury, replenished by market activity.