CopilotKit's new open-source Channels SDK lets teams run one AI agent natively across Slack, Microsoft Teams, and more — no rebuild required.
Most teams build an AI agent, wire it into a chat widget on their own site, and call it done — then discover the people who actually need it live in Slack or Microsoft Teams all day and never open that widget. CopilotKit's new open-source Channels SDK solves that mismatch directly: write your agent once, and run it natively inside every chat platform your team already uses. It launched this week and immediately picked up traction on both Hacker News and Product Hunt, landing on Product Hunt's daily leaderboard and the Hacker News front page the same day — a rare combination that usually signals a tool solving a problem a lot of builders actually have.
The Channels SDK is an MIT-licensed, open-source library that takes any agent built on the AG-UI protocol and deploys it into Slack, Microsoft Teams, and — soon — Discord, Telegram, and WhatsApp, from a single codebase. Instead of rebuilding your bot's UI and logic for every platform, you write "Channels JSX" once, and the SDK renders it as native Block Kit in Slack and Adaptive Cards in Microsoft Teams, with more platforms on the roadmap.
Under the hood, the agent keeps the capabilities that make it useful in the first place: it can stream responses, call tools and MCP servers, render interactive generative UI components, pause for human-in-the-loop approval before taking a risky action, accept file uploads, and carry memory with the user as they move between channels. Your agent logic and application code stay in your own infrastructure — CopilotKit's layer can be self-hosted or run through CopilotKit's managed intelligence layer for enterprise deployments.
It's also framework-agnostic by design. Teams already building on LangGraph, CrewAI, Mastra, Pydantic AI, or Google's ADK can plug straight in, since the SDK works with anything speaking the AG-UI protocol rather than locking builders into CopilotKit's own agent runtime.
For AI agencies and automation teams, the pitch is less about the technology and more about distribution. A support or ops agent is only as useful as the surface it lives on, and most businesses coordinate through Slack or Teams, not a bespoke web chat window. Historically, getting an agent into those tools meant a separate integration project per platform — different SDKs, different UI primitives, different approval flows — which made "just add it to Slack" a nontrivial line item on a proposal.
The Channels SDK collapses that into one build. An agency delivering a client's internal support bot, an approvals workflow, or a research assistant can ship it to Slack this month and Teams next month without touching the underlying agent. That's a meaningful efficiency gain for anyone billing hours to stand up AI workflows inside a client's existing stack rather than asking them to adopt a new app.
Because the SDK sits on top of AG-UI rather than replacing an agent framework, adoption doesn't require ripping out existing agent logic. A team with a LangGraph agent already handling a task can add the SDK as a delivery layer, define how its outputs should render as Channels JSX, and get native-feeling experiences in each platform — human-in-the-loop approval cards in Teams, interactive Block Kit messages in Slack — without maintaining separate codepaths per integration.
The human-in-the-loop approval piece is worth calling out specifically. Agencies deploying agents that can take real actions — sending emails, updating records, executing a workflow step — need a way for a human to sign off before anything ships, and having that pattern built into the SDK rather than hand-rolled per client is a meaningful head start.
This is most relevant to teams already building agents on an AG-UI-compatible framework and looking for faster ways to put those agents in front of end users where they already work. It's less relevant if you're not yet building agents at all, or if your deployment target is a standalone product rather than internal business chat tools. For agencies specifically, it's worth a look as a way to cut the cost of "and also make it work in Slack" from future proposals.
Our cutting-edge features simplify collaboration and creativity, making your workflow intuitive and efficient. Transform your vision into reality effortlessly with Hadidiz Flow.
