Cardstack Rewards

Executive Summary

AI code generation means you can ask for anything—and get working software in minutes. The unintended consequence: we ask for the same things over and over. Every developer burns tokens regenerating CRMs, dashboards, and workflows that thousands have built before.

Traditional sharing systems like GitHub weren't designed for this—they share entire source code repositories, not the granular artifacts that drive AI generation: system prompts and custom instructions, reusable components and data schemas, UI templates and styling presets, API integrations and tool definitions, even reference images and seed content.

What we need is not just easier sharing but a marketplace—where the artifacts you use for generation and the outputs you create can be published, discovered, and monetized. Creators earn when others build on their work. Value flows to human judgment instead of raw compute.

The solution requires a hybrid architecture that separates two kinds of costs:

• Material costs—AI inference, storage, compute, API calls—have real dollar-denominated expenses. For these, we use Boxel Credit: an off-chain metering system, efficient and stable.
• Alignment costs—the incentives that keep the ecosystem healthy: encouraging people to submit their work, enabling them to sell what they create, rewarding curation that surfaces quality, giving early participants a stake in what they're building. For alignment, we need the CARD token: a coordination mechanism with more flexibility than micropayments.

Cardstack Rewards is the name for this system—the flywheel connecting material costs to alignment incentives. Catalog growth drives transaction volume. Transactions drive burns and rewards. Rewards attract creators. Creators improve the catalog. Curation happens through staking: vouchers put CARD at risk when they endorse submissions, earning ongoing revenue when items succeed, losing stake when items fail.

Why Now: AI Changes the Economics of Software

For decades, businesses carved their workflows into silos—HR software here, CRM there, project management elsewhere—each charging $5 to $100 per user per month. The subscription model made sense when building software required teams of specialists working for months. You paid because you couldn't build it yourself.

That calculus has shifted. Anyone who has used Cursor, Claude Code, or Lovable knows the feeling: ask for a dashboard, a storefront, a workflow, and get working code in minutes. Once you've experienced that, paying fragmented subscriptions for dozens of siloed apps feels arbitrary.

AI-Generate-Everything Is Wasteful

But generating everything from scratch doesn't work either. Each generation burns tokens. Each risks bugs. Each payment flows to compute providers. The patterns are well-established—why regenerate them?

Remix and Pay the Creators

What makes sense is remix. Get proven solutions. Remix for your needs. Buyers get source plus license—full access to customize, extend, and deploy however you want.

The Catalog: Skip Generation, Remix Proven Solutions

Your AI searches the catalog before generating from scratch. Need a CRM? It finds proven templates that match. Browse what the community has built, or let AI find it for you—either way, you get full source to customize.

This is not an app store. You're not downloading finished products to use unchanged. You're acquiring starting points—proven solutions you adapt, extend, and transform. The Cardstack Catalog is infrastructure for remix.

Like npm, but for AI-Generated Software

The Catalog works like package managers developers already know—npm, pip, crates—but designed for the artifacts that drive AI generation. Builders submit their work. Submissions go through a vetting and review process powered by the staking mechanism. Approved components enter a commons where others can discover, remix, and build upon them.

The difference from traditional package managers: creators earn royalties when their work gets used. The difference from app stores: buyers get source code they can modify. The economics shift from consumption to contribution.

Discovery: Human and AI

Browse by category—Business, Creative, Productivity, Finance, E-commerce, AI Tools—or search by keyword. But the more powerful path: describe what you need, and the AI assistant searches the Catalog as part of its generation process. When proven solutions exist, it recommends them. When they don't, it generates.

This inverts the default behavior of AI coding tools. Instead of generating first and searching never, the assistant searches first and generates only what's missing. Remix before rebuild.

What You'll Find

Without a catalog, all value flows to the proprietary AI labs. Every regeneration is another payment to OpenAI, Anthropic, Google—compute providers who capture the upside while human expertise evaporates into chat logs. The Catalog changes where value goes. Instead of burning tokens to regenerate patterns, you pay creators for proven work.

Two Kinds of Costs: Material and Alignment

Running an AI software ecosystem has real costs. Inference, storage, compute, API calls—these are material costs with dollar-denominated expenses paid to external providers. Someone has to pay OpenAI, AWS, Anthropic.

For material costs, we use Boxel Credit: an off-chain metering system, efficient and stable. Like Roblox credits or an AWS balance, you top up, you spend, the math is simple.

But there's a second kind of cost: alignment costs. These are the incentives that shape ecosystem behavior—encouraging people to submit their work, enabling them to sell what they create, rewarding curation that surfaces quality, giving early participants a stake in what they're building.

CARD Token: Alignment via Rewards

Think of airline miles. When you fly, you pay the material cost—the ticket price. But on top of that, the airline attaches a rewards system: points that accumulate, that can be redeemed for upgrades, that grant status, that create loyalty.

CARD Token works the same way. It's not trying to replace dollars for material costs—Boxel Credit handles that. CARD is the alignment layer: the rewards program that coordinates ecosystem behavior beyond simple transactions.

Why a Token Instead of Points?

• Supply flexibility: Grow or shrink the supply based on ecosystem needs
• Transferability: Tokens move freely, creating liquidity and price discovery
• Partnerships: CARD can integrate with other platforms, creating network effects
• Governance: Token holders can participate in decisions
• Transparency: On-chain tokens are auditable

Membership Tiers: Status You Earn or Buy

CARD has the same structure as airline status. The token is fungible—interchangeable units you earn and spend. But membership tiers are more like NFTs—distinct status levels that unlock specific benefits.

How do you achieve tier status? Multiple paths: cumulative platform spending earns tier progression; locking CARD tokens guarantees tier status; publishing successful catalog items advances your tier; or simply pay directly for tier access. Early supporters lock in advantages that fade for latecomers.

What does status unlock? Access tiers (standard to full catalog including premium listings), earning multipliers (1× to 10×), catalog slots (limited to unlimited), discovery advantages (ranking, featured placement, search weighting). Higher status means more visibility—quality builders who demonstrate commitment surface more readily.

The Security Budget: Trust Through Staking

Quality determines the Catalog's value. Bad components—buggy code, security vulnerabilities, malicious behavior—destroy trust faster than good components build it. But centralized review doesn't scale and creates bottlenecks.

Cardstack Rewards solves this with staking. Builders submit components. Vouchers stake CARD on submissions they believe in, putting capital at risk. Submissions with sufficient vouching proceed to the Catalog. Vouchers earn ongoing revenue when items sell. If items prove harmful, vouchers lose their stake.

Professional Auditors

Smart contract security auditors—firms and independents who review blockchain code—can expand their market to Catalog submissions. Audit firms stake CARD to vouch for submissions they've reviewed, earning ongoing revenue from successful components. The same rigor applied to DeFi protocols applies to composable software.

Vouchers can also hire auditors to improve their odds—paying for professional review before committing stake. This creates a market for due diligence: auditors earn fees, vouchers reduce risk, and the Catalog benefits from higher-quality submissions reaching users.

Security Incident Response

Even with rigorous pre-submission vetting, issues emerge after components go live. Any user can file a security report. Credible high-severity reports trigger temporary blocks while under review. A panel of staked auditors, Foundation representatives, and domain experts evaluates evidence.

Foundation Support: Momentum Grants

AI generation is expensive—hundreds of dollars in inference costs to iterate, debug, and refine. Creativity and willingness to learn aren't correlated with wealth. Some of the most valuable potential contributors have skills and taste but not resources.

Cardstack Foundation takes a different approach: momentum grants. Instead of rewarding what you did, we amplify what you're doing. Think of an e-bike. You pedal, and the motor amplifies your effort—the more you pedal, the more power you get. Stop pedaling, the assist fades. The system rewards motion, not weight.

On-Chain Architecture

CARD Token exists on two layers. Ethereum mainnet holds the primary ERC-20 token, minted in 2018—the canonical asset for long-term holding, major transactions, cross-ecosystem portability. Base L2 holds bridged CARD for daily usage and staking—low-cost transactions, high-frequency operations.

Champer: The Revenue Management Protocol

Champer^† is a smart contract protocol for batch reward distribution based on activity analytics. Unlike off-chain accounting, Champer anchors all calculations on-chain—transparent, verifiable, trustless. The protocol is live on testnet (Gnosis Chain) and powers the first test run of the CARD Reward system.

Champer pairs the scalability of off-chain analytics with the trustlessness of on-chain verification. This hybrid approach solves a fundamental tension: complex reward calculations require computational power that would be prohibitively expensive on-chain, but final distributions must be transparent and tamper-proof.

^† Champer stands for: Consensus through the Harmony of Analytical Machines Producing Evaluated Results

Boxel Credit and CARD: Two Layers, One Economy

Two currencies serve different purposes. Boxel Credit—like dollars in your AWS account—provides stable pricing for metered services. No volatility when running inference. CARD Token—like airline miles that appreciate—captures ecosystem growth. When the catalog thrives, token holders benefit.

The currencies convert freely. Pay with CARD at a discount. Earn in CARD, withdraw as cash. Stake CARD for tier benefits. Daily operations stay simple; deeper mechanics—staking, vouching, governance—align long-term incentives.

Boxel Credit: The Operational Currency

Boxel Credit bridges the ecosystem to fiat-denominated infrastructure—AI inference, cloud services. No inference network today is denominated in crypto; conversion must happen. Boxel Credit handles this translation.

Acquire it through fiat purchase, CARD redemption at discounted rates, or earnings from catalog sales. The unified billing matters: one balance for everything, no separate accounts per service, no tracking across providers.

CARD Token: The Alignment Layer

CARD Token represents deeper ecosystem participation. Staking for tier status and vouching power. Earning through contribution, curation, and sales. Governance participation as the ecosystem matures. Long-term value capture as the ecosystem grows.

The two layers work together: Boxel Credit for operations (AI queries, hosting, catalog purchases), CARD Token for alignment (earnings, staking, vouching). Run inference with credit. Earn from sales in CARD. Stake CARD for tier benefits. Vouch with CARD at risk. Daily operations stay simple; economic alignment runs deep.

A Worked Example: Follow the Money

A user buys a $10 CRM template from the Catalog. Where does that $10 go? The creator receives ~75%—around $7.50—directly. The remaining ~$2.50 flows into the ecosystem: a portion covers operating costs, while the rest enters the recirculation pool. That pool purchases CARD from the open market, then splits it between burns (permanent supply reduction) and reward distributions to vouchers, purchasers, and affiliates.

The interactive diagram below traces every dollar through the system. Click to explore the full breakdown and adjust the parameters yourself.

The mechanics are deterministic. Burns are automatic—triggered by submission fees, slashing events, and certain transaction types. The flows are auditable through the proof-of-circulation ledger. Skeptics can verify that the economy operates as described: not promises, but queryable records.

At scale, these numbers compound. A thousand transactions monthly means $7,500 to creators, $1,050 in CARD burned, and $700 distributed through reward pools. The reinforcement is straightforward: more creators improve the catalog, a better catalog attracts more buyers, more transactions mean more burns and rewards, and rewards attract more creators.

The Seven Uses of CARD

CARD is a multi-purpose token designed for ecosystem coordination. Unlike single-purpose tokens, each use of CARD reinforces the others—creating a web of utility that makes the token integral to serious participation in the Boxel ecosystem.

Each use creates demand for the token while contributing to ecosystem health. Staking removes tokens from circulation. Burns permanently reduce supply. Rewards attract contributors. Discounts encourage spending. The seven uses form an interlocking system where participation in one reinforces the others.

Expanding the CARD Economy

The Catalog is open infrastructure—not tied to a single host. While Boxel.ai is the initial deployment, private instances, enterprise deployments, and third-party platforms can all tap into the same Catalog foundation. The economics of CARD extend across every deployment.

CARD utility extends beyond software. Physical products, consulting, training—anything participants offer can be priced in CARD. Any community can affiliate and offer tiered benefits: "We accept CARD" becomes a signal of ecosystem alignment, with membership tiers traveling across affiliated communities.

As more data sources connect—usage analytics, transaction records, contribution metrics—Champer analyzes both private and public datasets to shape optimal reward distributions. The vision: a membership as valuable as American Express Membership Rewards®, recognized across a network of partnerships, but built on-chain with Champer's analytics engine and AI coordinating discovery, promotions, and cross-ecosystem benefits.

The Mechanics of a Growing Ecosystem

Ecosystems either compound or decay. The difference is whether the mechanisms reinforce each other.

In this system, every catalog transaction triggers a burn. Not discretionary—automatic. Submission fees burn. Slashing burns. A portion of platform revenue converts to CARD and exits circulation permanently. Meanwhile, staking locks tokens for tier benefits and vouching power. Supply tightens from both directions.

Participation requires the token. Staking unlocks tier benefits and discovery advantages. Vouching requires locked CARD—curators need skin in the game. Catalog discounts favor CARD payment over fiat. The token isn't optional for serious participants; it's the mechanism of participation.

Conclusion: Why This Time Is Different

The Missing Energy

Most token flywheels never spin. The mechanics are sound—burns, staking, network effects—but the ecosystems stay inert. Web3 has spent years perfecting coordination mechanisms for communities that never reached critical mass. The virtuous cycles exist on paper.

In trading-oriented ecosystems, the loops feed on themselves: crypto users trading with crypto users, value circulating inside a closed system with no external injection. For web3 expanding beyond on-chain financial games into real software utility, there are few success stories.

AI changes the equation.

AI software is being adopted at unprecedented pace. Real users paying real money for real productivity gains—not speculation, not governance tokens, not promises of future utility. There's actual cash flow between users and AI applications, even if margins remain thin. The energy is external to crypto. It exists whether or not anyone cares about tokens.

Boxel is a pure-play AI application. A SaaS replacement platform that doesn't depend on web3 adoption to be useful. Users come for the AI workspace, the composable software, the remix economics. They stay because they build faster and ship more. The token layer is opt-in, not required.

This separation is the key. Revenue flowing through Boxel injects external energy into the flywheel. Not crypto trading with crypto—real economic activity from users who may never hold a wallet. Every catalog purchase, every AI inference call, every hosting fee feeds the recirculation mechanism. Burns happen. Rewards distribute. Supply tightens. And the source of that activity is a product people use for non-crypto reasons.

Build something valuable. Publish to the Catalog. Earn CARD. Stake for tier benefits. Vouch for quality. Earn more CARD. Spend on AI, hosting, components. Build something better. Repeat.

The test remains simple: do builders earn more here than elsewhere? For the first time, the answer can be yes—because a commons compounds. Every component makes the next one easier to build, richer in context, more discoverable. AI accelerates this: contributions become training data, recommendations, building blocks for the next generation. Your work appreciates as the catalog grows around it—and rewards from downstream value flow back to you.

What will you build—and what will you earn?