Report Abuse

Basic Information

AgentiPy is a Python toolkit for developers who want to enable AI agents to interact with blockchain applications, with a focus on Solana and Base. It provides a structured set of tools and examples to simplify building dApps and agent-driven workflows such as automated trading, token management, NFT minting and collection management, staking, liquidity operations, and DeFi interactions. The project exposes utilities for asynchronous blockchain operations, price and market data retrieval, wallet balance checks, transfers, and transaction confirmation. It also includes a built-in Model Context Protocol (MCP) module and an MCP server implementation so language models and external agents (for example Claude Desktop) can invoke standardized on-chain actions. The README contains installation steps, security guidance about private key handling, quick start code samples for transfers, balance queries and swaps, and guidance for integrating with Langchain for natural language command interpretation and dynamic workflow generation.

Links

App Details

Features
AgentiPy ships a comprehensive toolset and integrations intended for programmatic on-chain activity. It documents broad protocol support on Solana (examples include Jupiter, Raydium, Metaplex, Helius, MoonShot, Orca, OpenBook, Pyth, and many others) and mentions Base support. The toolkit emphasizes asynchronous operations for efficient RPC use and includes utilities such as price fetching, balance fetchers, token transfer and deployment managers, trade and staking managers, NFT minting tools, RugCheck reports, and CoinGecko market data integration. It is designed to be extensible so developers can add custom protocols and actions. The repo includes an MCP layer with ALL_ACTIONS aggregation and an MCP server to register tools and expose them to LLM-driven workflows. Langchain integration examples show how to map natural language instructions to tool calls.
Use Cases
AgentiPy helps developers and AI researchers quickly prototype and operate agent-driven blockchain workflows by providing prebuilt actions, protocol adapters, and an MCP interface to standardize tool invocation. It reduces integration overhead for common tasks like SOL/SPL transfers, token swaps on Jupiter, balance checks, NFT minting via Metaplex, and obtaining market metrics from CoinGecko or price feeds like Pyth. Langchain examples illustrate how language models can parse user intent and call AgentiPy tools, enabling natural language trading and automated DeFi strategies. The MCP server enables seamless exposure of on-chain actions to external agents such as Claude Desktop, facilitating standardized orchestration. The README also provides security guidance, recommended prerequisites, and concrete quick start examples to lower the barrier to safe testing on devnet/testnet before mainnet deployment.

Please fill the required fields*