Build Full-stack Agents with CrewAI & CoAgents
...with fully open-source tooling.
Today, we have reliable tools to create powerful agentic workflows.
Yet, the main challenge is transitioning Agents from local setups to user-facing apps.
How do you embed a seamless UI?
How do you enable user control when necessary?
How do you unify the Agent’s state and actions with the app?
CoAgents, an open-source framework by CopilotKit, is solving this.
It is the only framework that offers all the essential Agentic front-end building blocks to set up full-stack agentic apps.
Most recently, they extended full-stack development support to Agents built with CrewAI.
This includes both regular Agentic Crews and Flows.
This way, you can give your CrewAI Agents a front-end brain—so they don’t just work behind the scenes but rather become an integral part of the user experience.
Let’s learn more about it today!
But before we proceed…
The CopilotKit team is organizing a webinar tomorrow (28th March) on helping you build full-stack agentic apps.
You will learn:
Building Full-stack Agentic apps with CrewAI and CopilotKit.
Using CopilotKit to build UIs for multi-agent workflows.
Integrating CopilotKit into CrewAI Agentic stack.
Going from zero to agentic UI in minutes.
You can register for free here →
Now let’s get back into the newsletter issue!
The problem
Whenever an Agent needs to perform a complex task with multiple steps, one mistake can derail the entire operation.
To avoid this, they need feedback mechanisms, such as human-in-the-loop (HITL), where a human can steer the Agent as needed.
Not only this, it is also desired to:
Chat/interact with the Agent.
Generate UI components on the fly.
Know what the Agent is doing at each stage.
And building this from the ground up is tough.
The solution—CoAgents
CoAgents is an open-source project by CopilotKit that provides the infrastructure to build such workflows.
This visual depicts the four major capabilities of CoAgents to build full-stack AI Agents:
Chat → You can integrate chat functionalities into your Agentic applications.
Shared state → Your application can see everything the Agent is doing, and the Agent can see everything that happens inside the application.
Agentic UI → Render the Agent’s real-time state so the user can see what the Agent is doing.
Human-in-the-loop → Specify breakpoints in the agent process at which Human input/approval is required. This allows for greater safety as well as better Agent performance.
Until we get to fully autonomous AI Agents, they will be driven by human-in-the-loop workflows.
Learning these skills will put you towards an exceptional career in AI if you intend to build one.
To help you do that, the CopilotKit team is organizing a webinar tomorrow (28th March).
You will learn:
Building Full-stack Agentic apps with CrewAI and CopilotKit.
Using CopilotKit to build UIs for multi-agent workflows.
Integrating CopilotKit into CrewAI Agentic stack.
Going from zero to agentic UI in minutes.
You can register for free here →
Do star CopilotKit repo to support their work: CopilotKit GitHub.
🙌 Also, a big thanks to the CopilotKit team for working with us on today’s issue and for being a long-time supporter of this newsletter.
👉 Over to you: What problems will you use CopilotKit for?
Thanks for reading!