project screenshot 1
project screenshot 2
project screenshot 3
project screenshot 4

CrewKit

CrewKit integrates CrewAI with AgentKit to unlock collaborative multi-agent workflows for onchain operations. Enjoy enhanced wallet management, multi-network support, and a ready-to-use chatbot for seamless crypto interactions—all in one package.

CrewKit

Created At

Agentic Ethereum

Winner of

Coinbase

Coinbase Developer Platform - Best AgentKit Contributions 3rd place

Coinbase

Coinbase Developer Platform - AgentKit Pool Prize

Prize Pool

Project Description

CrewKit is a new extension to AgentKit that brings CrewAI’s collaborative multi-agent framework into the mix. By adding the dedicated cdp-crewai package (modeled after the existing LangChain integration), CrewKit enables developers to build decentralized applications where agents work together on complex onchain tasks. Key features include:

Collaborative Agent Workflows: Empower multiple agents to work in concert, leveraging CrewAI’s strengths over single-agent setups. Comprehensive Toolkit: Full support for AgentKit core actions, alongside enhanced wallet management (via persistent MPC wallets) and multi-network connectivity. Gasless Transactions: Out-of-the-box support for gasless operations on popular networks for tokens like USDC, EURC, and cbBTC. Real-world Example: A chatbot example that demonstrates how to use CrewAI-powered agents to execute seamless crypto transactions. This integration expands AgentKit’s versatility while remaining fully backward compatible—users can continue with the stable LangChain integration if desired.

How it's Made

CrewKit was developed by creating a new Python package, cdp-crewai, which mirrors the design and functionality of the cdp-langchain package. Here’s how we did it:

Architecture: Built on top of the core primitives from cdp-agentkit-core, the package introduces modules that handle CrewAI integration, wallet management, and network support. Modular Design: The new functionality is encapsulated entirely within the cdp-crewai package, ensuring no disruption to existing AgentKit modules. This design allows for independent updates or unpublishing if needed. Testing: We implemented unit tests and conducted end-to-end testing using a provided chatbot example, following the established testing patterns in AgentKit’s contributing guidelines. Technologies: The project leverages Python 3.10+, CrewAI, LangChain, and other standard libraries to ensure robust onchain interactions and agent collaboration. Best Practices: We adhered to the AgentKit contribution standards, ensuring that CrewKit’s addition is both seamless and easily adoptable by the community.

background image mobile

Join the mailing list

Get the latest news and updates