Project Name: Stormrae
Project Brief: Stormrae is the participation layer for AI development. We transform essential AI data tasks (adversarial testing, labeling, preference evaluation, validation) into engaging consumer experiences. We are building participatory AI infrastructure.
The Problem: AI labs spend $65B yearly on evaluation and data labeling with massive inefficiency. Current processes are expensive, closed, and fail to generate diverse, high-quality results. Web3 players have failed to capture this market because they lack engaging user experiences and verifiable data pipelines.
The Market Shift: The AI industry is pivoting. Budgets are moving from new model releases to fine-tuning and improvement. Labs need continuous red-teaming data, adversarial testing, and safety benchmarks to make existing models better. Stormrae is built for this moment, enabling incentivized public contribution at scale.
The Solution: We build consumer-to-enterprise infrastructure for AI data tasks. Users compete, label, and evaluate through gamified experiences with real stakes. Every interaction generates enterprise-grade datasets.
Target Market: AI labs needing scalable evaluation, enterprises deploying AI solutions, and data providers seeking new monetization channels.
Uniqueness: We are building the infrastructure for participatory AI development. We are building and releasing multiple consumer experiences masking participation in AI development while allowing normal end users to actually be involved in the AI development process.
How will you integrate 0G? Stormrae positions 0G as the main player in web3 AI development. We are looking to store our entire dataset using 0G decentralized storage solution.
Technical Architecture & Implementation Plan
Stormrae is built on a hub-and-spoke architecture where multiple client applications interface with a central API Gateway that orchestrates all operations.
Core Components:
-
Client Layer: dApps, mobile apps (Seeker), and admin dashboard
-
API Gateway: Central orchestrator handling authentication, data management, and service coordination
-
AI Layer: Autonomous AI agents using LangChain and LangGraph for stateful response generation
-
Blockchain Layer: Smart contracts for prize pools, credit purchases, and reward distribution
Key Technical Features:
-
Asynchronous job processing for AI generation to maintain UI responsiveness
-
Multi-layered security with input validation and security classifiers
-
Horizontal scalability for AI worker clusters based on real-time demand
-
Real-time interaction layer for low-latency chat, level progression, and instant game feedback
-
User profiling & segmentation layer based on in-game behavior and on-chain activity, used to personalize limits, rewards, and matchmaking
-
Multiple anti-fraud systems, including detection of AI-generated user inputs to prevent automated LLM-to-LLM attacks and abuse
-
Trust scoring system based on in-game behavior to identify suspicious patterns, sybil-like activity, and coordinated attempts
-
Event-sourced game architecture where all interactions are logged as events, paired with a central state machine that controls game mechanics and difficulty, including dynamic rules influenced by real-world signals (e.g. external weather conditions)
-
Deterministic level progression engine (Levels 1–5) with increasing difficulty, configurable win conditions, and prize eligibility enforcement
Integration with 0G Infrastructure
Stormrae will use 0G Storage as the foundation for transparent, auditable AI evaluation.
0G Storage Integration:
We will anchor cryptographic hashes of user interactions, AI agent responses, and adversarial attempt outcomes to 0G Storage, creating an immutable audit trail for AI evaluation. Every prompt, every exploit attempt, every successful jailbreak becomes verifiable on-chain.
While proprietary AI models and detailed user prompts remain off-chain for privacy and performance, we store proof-of-interaction hashes on 0G. This allows enterprises to verify data integrity without exposing sensitive information.
What Gets Stored on 0G:
-
Cryptographic hashes of user prompts and AI responses
-
Timestamps and user wallet addresses for each interaction
-
Level progression records and achievement unlocks
-
static assets such as images, css, js etc
-
private encrypted data used as backup and auditability
Benefits for 0G:
-
Continuous storage demand from consumer-scale AI evaluation activities
-
High-throughput verifiability for enterprises needing auditable AI testing
-
Showcase of 0G’s capability to support real-time, interactive AI applications
-
Establishes 0G as the trusted infrastructure for transparent AI evaluation