Architecture

Three-Layer Architecture

Wildmeta's architecture consists of three distinct layers that work in harmony to create a comprehensive agentic trading ecosystem. Each layer serves a specific purpose while seamlessly integrating with the others through the x402 protocol foundation.

Layer 1: Skills Marketplace - Powers the Data and Tools

The Skills Marketplace forms the foundational data and capability layer of Wildmeta's ecosystem. This open marketplace enables various skill providers to offer specialized trading capabilities, analytical tools, and data services through standardized x402 interfaces.

Trading Skills Available

The marketplace hosts diverse trading capabilities ranging from analytical tools to fully autonomous strategies. Trend analysis and position analysis services, provided by partners like nof1.ai, offer sophisticated market insights. Algorithmic trading strategies encompass pure market making, directional trading, and volume-based approaches. Quantitative trading strategies leverage advanced models such as BTC-baseline-1h-LSTM for systematic trading approaches.

Additional services include comprehensive backtesting capabilities, low-risk asset management through HLP and LLP integration, and sophisticated copy trading systems that can mirror individual traders or curated lists of traders with specific configurations and risk parameters.

Data Provider Ecosystem

Skills within the marketplace can select from multiple data provider options, creating competitive dynamics that benefit end users. Market data sources include direct Hyperliquid data through Hyperliquid RPC endpoints and comprehensive liquidation data from CoinGlass. Post-processed intelligence includes Wildmeta's proprietary trader profile data and trade story analysis.

The platform also provides access to exclusive private data sources, including 24-hour biggest position records on Hyperliquid, offering information asymmetries that can provide trading advantages. This multi-sourced approach ensures data reliability while maintaining competitive pricing through provider diversity.

Layer 2: LLM Orchestration - Thinks

The LLM Orchestration layer serves as the reasoning and execution brain of the entire system. This sophisticated engine interprets user intents expressed in natural language and connects them with appropriate Model Control Protocols (MCPs) to execute complex trading operations.

Natural Language Processing

The orchestration engine processes complex trading instructions such as "Create a market-making strategy on ETH-USDT," "Follow this trader's positions," or "Rebalance portfolio and put 50% of profit into HLP weekly." This natural language interface eliminates the need for users to understand complex trading interfaces or programming languages.

Technical Infrastructure

The system operates on a self-deployed LLM based on ChatGPT-OSS running on dedicated H100 GPU infrastructure. This ensures low latency, high availability, and the ability to handle complex reasoning tasks required for sophisticated trading decisions. The infrastructure provides the computational power necessary for real-time market analysis and strategy optimization.

MCP Integration

Through Model Control Protocol integration, the LLM layer seamlessly connects user intentions with available skills and data providers. This creates a highly composable system where different capabilities can be combined dynamically based on market conditions and user preferences.

Layer 3: Execution Gateway - Acts

The Execution Gateway represents the action layer where trading decisions translate into actual market operations. This layer facilitates and monetizes on-chain execution while maintaining connections across multiple decentralized exchanges.

Multi-DEX Connectivity

The gateway maintains connections to various DEXs including Hyperliquid, Aster, EdgeX, and other emerging platforms. This multi-venue approach ensures optimal execution prices and reduces dependency on any single platform while providing access to diverse liquidity pools.

x402 Completion

The Execution Gateway completes the x402 protocol implementation by addressing the unique challenges of trading where requests and executions don't maintain a one-to-one relationship. AI analysis doesn't necessarily translate to immediate actions, and bot executions can't be predicted in advance. The gateway implements revenue sharing by confirmation, ensuring fair compensation for successful executions.

Revenue Model Integration

The gateway implements a sophisticated fee structure that includes execution fees charged via x402, creating sustainable revenue while maintaining competitive pricing. Protocol revenue contributes to $WILD token buybacks, creating value for token holders while sustaining platform development.

Integration and Synergy

The three layers work together seamlessly: Layer 1 powers the data and tools, Layer 2 thinks, and Layer 3 acts. Wildmeta connects them all — and x402 makes it smooth.

This architecture enables unprecedented composability where users can combine different analytical tools, trading strategies, and execution methods without technical complexity. The x402 protocol ensures transparent pricing and autonomous payments throughout the system, creating an efficient marketplace where value flows to the most effective providers.

The modular design allows for continuous innovation as new skills and data providers can integrate without disrupting existing functionality. This creates a robust ecosystem that evolves with market needs while maintaining stability for existing users.