Wayfinder
Wayfinder is an omni-chain AI tool that enables autonomous AI agents to navigate and interact with blockchain environments through dedicated Web3 wallets they control. The platform allows both technical and non-technical users to direct AI agents to perform complex blockchain operations across multiple networks. [1] [2] [3] [4]
Overview
Wayfinder serves as a foundational infrastructure that bridges artificial intelligence and blockchain technology, creating a system where AI agents (called "shells") can autonomously navigate blockchain ecosystems. These shells can control digital wallets, execute transactions, and interact with smart contracts across multiple blockchain networks. The platform addresses several key challenges in the blockchain space, particularly the complexity of cross-chain operations and the technical barriers that prevent widespread adoption.
At its core, Wayfinder functions through what it terms "wayfinding paths" - essentially mapped routes that guide AI agents through blockchain environments to reach specific endpoints such as decentralized exchanges, NFT marketplaces, or other Web3 applications. These paths collectively form an ecosystem graph that serves as a comprehensive map of blockchain networks, allowing shells to navigate efficiently between different protocols and applications.
The platform is designed to be accessible to users regardless of their technical expertise. Through natural language interfaces powered by Large Language Models (LLMs), users can direct their shells using conversational prompts rather than complex code. This approach significantly reduces the technical barriers to blockchain interaction, potentially expanding the accessibility of Web3 technologies to a broader audience.
Wayfinder's architecture is chain-agnostic, allowing it to operate across various blockchain ecosystems including Ethereum, Solana, Base, and Cosmos. The platform is built on Solana but designed to work seamlessly with multiple networks, prioritizing high-speed transactions necessary for AI agent operations. [1] [2] [3] [4] [8] [5]
Technology
The technological foundation of Wayfinder centers on its wayfinding paths and ecosystem graph. These components enable AI agents to navigate blockchain environments with precision and efficiency:
- Wayfinding Paths: These are instruction sets that guide shells to specific endpoints within blockchain ecosystems. Each path contains the necessary information for an AI agent to navigate to a particular application or smart contract and execute specific functions .
- Ecosystem Graph: The collection of wayfinding paths forms a comprehensive graph that maps relationships between different blockchain contract addresses, assets, applications, and wallets. This graph enables shells to evaluate alternative routes to achieve specific objectives and select optimal paths based on factors like fees, security, and historical success rates .
The graph consists of various node types:
- Protocols (sets of applications, contracts, and assets);
- Contracts (on-chain logic with unique addresses);
- Contract Standards (ERC-20, ERC-721, etc.);
- Assets (tokens and NFTs);
- Functions (executable code segments);
- API Functions (protocol-level information providers);
- Routines (sequences of functions for specific tasks);
AI Shells
Wayfinder's AI agents, known as shells, are autonomous entities that can control Web3 wallets and execute blockchain transactions. Key features include:
- Memory Retention: Shells can write and read memories, allowing them to learn from past experiences and improve their decision-making over time. These memories include conversation memories from user interactions, tool memories for specific tasks, and long-term contextual memories from accumulated experiences .
- Blockchain Payload Awareness: Shells can incorporate blockchain data into their operations, enabling more informed decision-making based on current market conditions and blockchain states .
- Private Key Management: A critical innovation of Wayfinder is its approach to secure private key management, allowing shells to control Web3 wallets and request signatures from associated private keys while maintaining security .
- Explicit Intent: Shells operate within the bounds of explicit instructions from their owners. While they can execute complex operations, they cannot autonomously manage portfolios or transmit value at their own discretion without express direction .
Purpose-Built LLMs
Wayfinder leverages and enhances open-source Large Language Models (LLMs) to create specialized "expert" models for different types of shells. These purpose-built LLMs are designed to be more efficient and cost-effective than larger, closed-source alternatives. For example, the platform has developed a compact LLM specifically for Ethereum Virtual Machine (EVM) function calls that integrates with the wayfinding graph.
The platform is committed to open-source principles, with plans to open-source both the fine-tuning datasets and the weights of these models, enabling other developers to build upon this work. [1] [2] [3] [4] [6] [9]
Wayfinder Token (PROMPT)
Wayfinder's native cryptocurrency, PROMPT, serves multiple functions within the ecosystem:
- Shell Acquisition: Users spend PROMPT tokens to purchase and create AI shells.
- Network Operations: PROMPT functions as a gas token within the network, with fees being shared between wayfinding path developers and the protocol.
- Staking for Path Creation: Users stake PROMPT tokens to propose new wayfinding paths. These tokens remain staked as long as the path is active and are subject to slashing if the path is found to be incorrect or malicious.
- Governance: PROMPT token holders can participate in network governance, voting on ecosystem direction, staking and fee levels, elections, and reward parameters.
The token has a maximum supply of 1 billion tokens, with the following proposed distribution:
- 45% to holders of ecosystem tokens who cache their tokens within Wayfinder;
- 5% allocated to wayfinding rewards;
- Remaining allocations for investors, team, and other ecosystem participants.
Users can earn PROMPT tokens through various mechanisms:
- Receiving a percentage of throughput fees on wayfinding paths they create;
- Fulfilling bounties for developing requested wayfinding paths;
- Operating Verification Agents that identify errors in the network;
- Participating in additional community programs.
Governance
Wayfinder's governance model has several distinctive features:
- Token Holder Governance: PROMPT token holders can vote on network parameters, ecosystem direction, and other governance matters.
- AI Agent Participation: A unique aspect of Wayfinder's governance is that AI shells can participate alongside human stakeholders. Since shells hold PROMPT tokens to transact on the network, they can both vote and submit governance proposals. However, shells are limited to an aggregate total voting weight of approximately 20% to mitigate potential sybil attacks.
- AI Constitution: The platform includes an "AI Constitution" that establishes overarching principles guiding the behavior of the network's AI agents. This constitution defines actions shells are forbidden from pursuing and includes provisions for pausing all shells in case of concerning activities.
The governance process also involves monitoring and revising the methods by which wayfinding paths are evaluated and ranked, ensuring the network's continued safe and efficient operation as blockchain ecosystems evolve.
Use Cases
Wayfinder's technology enables various applications across different domains:
- NFT Minting: Users can train shells to become NFT-minting experts, handling the process of ingesting media, determining optimal formats, evaluating blockchain ecosystems, and overseeing the minting process.
- Trading and Portfolio Management: Shells can be deployed for various trading and investment management strategies, from simple dollar cost averaging to more complex applications like mirroring specific time-delineated volume weighted average price targets.
- Gaming: Wayfinder provides a framework for gaming applications, particularly in autonomous world ecosystems where it can enable a generative substrate of scarce resources.
- Customer Service: Shells can be trained with company documentation and prior communications to become subject-matter experts capable of handling customer support inquiries.
- Smart Contract Analysis: Users can teach shells to analyze new smart contracts and filter for appealing assets based on specific criteria. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]