SpyNet

Demo

About

The problem it solves ## The Problem It Solves SpyNet was created to solve critical structural and economic bottlenecks in the emerging **multi-agent AI economy**. As artificial intelligence evolves into networks of specialized, autonomous agents capable of collaborating and transacting at machine speed, the current digital infrastructure—especially in terms of **payments, traceability, trust, and discoverability**—is fundamentally inadequate. Despite a trillion-dollar opportunity on the horizon, the AI agent market today suffers from fragmentation, lack of standards, and operational friction. SpyNet introduces a unified solution by offering a **decentralized, blockchain-powered marketplace** for AI agents—enabling seamless creation, commercialization, discovery, and usage of agents under a trustless, auditable, and scalable framework. --- ### What People Can Use It For #### For Agent Creators Creators of AI agents—developers, startups, researchers, hackers, and students—can use SpyNet to: * **Monetize their agents natively via blockchain micropayments** SpyNet allows creators to set granular usage-based pricing (e.g., per API call). Traditional systems can’t process high-frequency, low-value transactions efficiently. SpyNet’s on-chain micro-payment infrastructure makes sub-cent monetization viable. * **Sell agents directly without intermediaries** Smart contracts replace traditional licensing, invoicing, or third-party integrations, enabling autonomous transactions between users and agents. * **Receive payments instantly in cryptocurrency**, eliminating delays, cross-border friction, and overhead fees from banks or platforms. * **Gain visibility and public reputation** Agent performance, usage logs, and user feedback are stored transparently, giving creators clear insights into what’s working and where to improve. * **Prove the value of their agents in a verifiable way** On-chain proof-of-work and usage history allow creators to demonstrate effectiveness without relying on unverifiable claims. #### For Agent Buyers Buyers of agents—including startups, SMEs, large enterprises, freelancers, and product teams—can use SpyNet to: * **Find reliable, specialized AI agents for specific problems** Through categorized listings and search by domain (e.g., legal, marketing, healthcare), buyers can identify agents that match their operational needs. * **Pay only for what they use, when they use it** Traditional licensing often requires upfront payments or subscriptions. SpyNet enables a *pay-per-contribution* model, where buyers pay per task or API call—no more paying \$50 for a \$0.05 task. * **Ensure accountability with verifiable, auditable logs** Every agent interaction is recorded on-chain, enabling buyers to audit performance, resolve disputes, and build trust. * **Avoid financial risks associated with non-performing agents** The system allows traceability and flexible usage, reducing the need for risky commitments to unproven technologies. * **Integrate agents into automated workflows** With API keys, public endpoints, and smart contracts, agents can be plugged directly into existing systems for real-time, autonomous execution. --- ### How It Makes Tasks Easier, Safer, or More Efficient #### 1. **Eliminates the Infrastructural Friction** SpyNet provides native tools for identity, payment, discovery, and access control. Creators don’t need to build or manage billing systems, contract flows, or dashboards—SpyNet handles this infrastructure. #### 2. **Enables True Machine-to-Machine Commerce** With autonomous micropayments, agents can be composed, chained, or orchestrated in real time—enabling complex, decentralized multi-agent systems (MAS) to function economically at scale. #### 3. **Reduces Risk for All Stakeholders** Through transparent usage logs, smart contract-enforced terms, and blockchain-based audit trails, every interaction is secure, verifiable, and immutable—reducing fraud and boosting trust on both sides. #### 4. **Accelerates the Maturity of the AI Agent Economy** By solving monetization, discoverability, and auditability, SpyNet accelerates adoption and unlocks value trapped in underutilized or undiscovered agents. --- ## Summary SpyNet solves the core problems facing both creators and users of AI agents: * No scalable way to monetize agents for granular tasks * No unified platform for discovery and quality verification * No viable economic model for micropayments * No transparent performance history for trust or accountability * No plug-and-play framework for real-world AI deployment

How it's made

Challenges we ran into ## Challenges I Ran Into One of the most complex challenges we faced during the development of SpyNet was implementing a **robust API usage monitoring system** that could **trigger accurate billing events via Chainlink** based on real-time consumption. ### The Problem We needed to design a system where: * Every time an **API endpoint from an AI agent was called**, that request needed to be: * Tracked with usage metadata, * Verified for legitimacy and origin (authenticated), * Mapped to a contract and associated user, * Converted into a verifiable billing event. * These usage logs had to **interact with Chainlink** to: * Generate cryptographically-verifiable **Proofs of Work**, * Trigger **on-chain billing mechanisms** (e.g., token transfers or credit deductions), * Record each interaction immutably for auditability. This required a delicate bridge between **Web2 (Node.js + Express backend)** and **Web3 (EVM-compatible smart contracts + Chainlink Functions)**, which introduced asynchronous communication, signature validation, and oracle latency concerns. ### How We Solved It 1. **Reverse Proxy Handler** We created a `proxyAgentCall` controller that intercepted every incoming request to an agent's endpoint. This handler: * Validated the contract and API key associated with the request, * Forwarded the request to the actual agent, * Captured the response and stored it alongside metadata (timestamp, payload, requester). 2. **Usage Log & Hashing** Every call generated a `usageHash` using a secure hash function (`keccak256`) that encoded the key interaction data. This acted as a **unique fingerprint** of the API usage event. 3. **Chainlink Triggering** We integrated **Chainlink Functions** to periodically fetch the usage logs or respond to specific usage events, and: * Verify the proof hash, * Match it to the contract terms, * Trigger an on-chain function to deduct credits or record usage. 4. **Smart Contract Adjustments** We had to structure our smart contracts (e.g., `NFTMarketplace`, `ContractRegistry`, `NetToken`) to: * Maintain mappings between users, agents, and credits purchased, * Emit events upon consumption, * Enforce limits and prevent overuse. 5. **Debugging the Asynchronous Pipeline** One major debugging hurdle was **ensuring data consistency across asynchronous steps**—especially when the Chainlink node queried the backend while the database was still writing the usage logs. We implemented internal retries and status checks to mitigate race conditions.