Start here—Bring external agents into Front!

👋 Welcome! If you've already built — or are thinking about building — your own AI agent and want it to work inside Front, this community group is for you. This post is your home base: what we're building, what we're asking of you, and how to get started.

What is BYOAgent?

Today, plugging your own agent into Front means a webhook, a human account standing in as the "agent," and a handful of API calls per reply. It works, but your agent isn't a real participant — there's no agent identity, no clean handoff to a human, no signal when someone takes over.

BYOAgent makes your external agent a first-class teammate in Front. It gets assigned conversations through your normal rules, replies under its own name and avatar, hands off to humans cleanly, can get @ mentioned in a conversation, and operates under the same governance your team already trusts — all without you running a fragile integration in the middle.

The promise: Bring any agent. Front makes it a teammate.

How to participate

We are currently in the early stages of development and are looking to better understand your use cases. At this time, the beta is open to all customers, but you will need to have an AI agent operational before you can test the framework.

    Pre-requisites:

• Already have an operational AI agent hosted in the tool or framework of your choice
• Access to the technical or developer resources required to handle webhook events and establish connections to MCP servers and APIs. Connecting an agent means hosting a server that accomplishes these actions either on your own infrastructure or on a platform like n8n or Zapier.

How to leave feedback
After joining, please leave feedback as a new post in this group.

We also have a product feedback survey that we’re paying super close attention to for shaping our upcoming priorities, so please take a moment to fill it out, whether you end up liking the framework or struggling with it!

What’s available now