← All specs
Phase 2-3 draft

Stoka Token — Economics

The economic layer. Earn Stoka by contributing (bug reports, beta tests, verified accounts, compute rental). Spend Stoka on platform compute and ecosystem apps. Stake Stoka on app success for NFT achievements and future revenue. Tied to compute price; designed so staking feels like investing in the ecosystem.

Stoka Token — Economics

Scaffold from raw notes. The mechanic is clear; the legal and UX specifics need iteration.

Thesis

Stake > subscribe. Subscriptions are arbitrary monthly prices divorced from contribution. Staking lets the contributor see the threshold at which their activity is "free," because the ecosystem's health covers it. Patch notes become the thing your stake funds.

The token ties three flows from ethos.md — compute, creativity, execution — into a single ledger that rewards contribution in kind.

The earn side

viz:kind-cardsplanned

Kind cards — component not yet built. Data parsed; rendering deferred.

raw yaml
kinds:
  - id: bug-reports
    name: Bug reports
    description: Rate-limited per day per verified account. Payouts weighted by quality (reproducibility, novelty, severity). Anti-spam requires phone verification.
  - id: beta-testing
    name: Beta testing
    description: Contract-bounded testing sessions for staked apps. Reward paid on completion with quality review. App pays out of its stake pool.
  - id: compute-rental
    name: Compute rental
    description: Rent your idle GPU(s) to the Stoka scheduler. Inbound jobs (training, inference, batch) pay market rate in Stoka, priced against raw compute cost.
  - id: creative-labeling
    name: Creative labeling
    description: Artifact tagging, moderation-adjacent judgment, Bon Appétit framing review. Smaller flat payouts; scales with trust ELO.
  - id: deployment-rebates
    name: Deployment rebates
    description: Apps deployed through Stoka (see trustless-deployment.md) receive immediate rebate on their first compute spend — potentially paid in excess, if the training they enable improves the system.

Anti-fraud: phone verification + daily caps + quality review. High-signal contributions (reproducible bug, novel prompt recipe accepted into Stoka's library) earn more than baseline. Trust ELO scales permitted throughput.

The spend side

  • Compute. Every Stoka-provided inference, training, or RAG query is priced in Stoka against the underlying compute cost. Transparent pricing — users see the exact conversion rate when they run a job.
  • Ecosystem apps. Apps deployed on the platform can price in Stoka. Users pay from their wallet; apps receive it into theirs.
  • Beta test entry. Staking Stoka buys you access to beta cohorts and gated drops.
  • Premium features in core platform. Graduated surfaces (larger artifact quotas, priority scheduling) unlocked by token balance or stake duration.

Tied to the price of compute. The protocol targets a floating conversion: one Stoka buys a fixed unit of compute-time across a reference GPU. When GPUs get cheaper, one Stoka buys more. When demand surges, token demand rises. The token is therefore a claim on future compute — the one thing every Stoka user will want.

The stake side

Staking is the primary capital action.

  • Stake on apps. An app deployed through Stoka declares a stake requirement. Users can lock Stoka behind the app's success — if the app hits milestones (acquisition, retention, revenue), stakers earn proportional returns. If it fails, stake unlocks minus any spent compute.
  • Stake on the platform itself. Platform stakers share in protocol fee flow and receive governance weight proportional to duration (long-dated stake > spot).
  • NFT achievements. Stakers on breakthrough apps/contributions get on-chain badges — "staked pre-traction on app X, early Stoka", etc. Tradeable but meaningful as provenance.
  • Visible coverage threshold. The UI shows the user exactly when their stake's yield covers their consumption. Crossing that line is the goal — the product incentive is self-sustaining membership, not monthly bleed.

Compute marketplace (Phase 3)

The GPU rental mechanism turns the platform into a two-sided market:

  • Supply. Individuals and orgs attach idle GPUs to Stoka's scheduler.
  • Demand. Apps, training jobs, and power users buy compute in Stoka.
  • Price oracle. The platform sets reference prices by GPU class; market discovers surcharges.
  • Settlement. Instant rebate to the supplier, held in Stoka.

This flips the monolith pattern (you pay API rates; the provider collects). Here, compute providers are paid directly, and the platform takes a thin coordination fee.

Regulatory surface

Inception relationship to existing specs

  • artifact-platform.md — artifact capture and usage generate on-chain provenance that the token can price.
  • messaging.md — pseudonym primacy lets contribution be scored without identity surveillance.
  • trustless-deployment.md — the deployment substrate through which apps enter the stake market.
  • stoka-bot.md — Stoka surfaces the economic state to the user (balance, coverage, staked positions) in-voice.

Open questions (TBDs)

  1. Token vs credit. Can Phase 2 ship with account credits (non-transferable, no legal exposure) and Phase 3 upgrade to a token? The economic intent works either way.
  2. Stake-on-app failure mode. If an app dies without milestones, do stakers get principal back, or are they writing options? Principal-protected stake is UX-friendly but caps upside; at-risk stake is more VC-like.
  3. Compute price oracle source. Market rate from whom? Crowd-sourced from GPU renters? Hardcoded by Stoka to a reference GPU class?
  4. Governance weight formula. Stake × duration × contribution-ELO? Needs modeling.
  5. Anti-Sybil under anonymity. Phone verification is weak. ZK identity primitives (BrightID, WorldID-style without the bio) are the likely path — TBD after Phase 2.

Related specs