SandboxAQ's open-source Switch lets Claude, LangChain, OpenAI, and Bedrock agents join your team's existing Slack, Teams, or Discord channels.
Every automation-minded team eventually runs into the same problem: you build a great AI agent, and then you have to decide where it lives. Slack bot? Standalone dashboard? Yet another browser tab nobody opens? SandboxAQ's new open-source tool, Switch, takes a different approach — instead of giving your agents a new home, it brings them into the team chat tools your company already uses: Slack, Microsoft Teams, and Discord.
Switch launched on GitHub in late August 2026 and hit Product Hunt's "Top Products Launching Today" list on September 9, ranking #1 with over 400 upvotes — a strong signal that the "stop building yet another agent UI" pitch is resonating with builders.
Switch turns an existing Slack, Teams, or Discord channel into what its creators call a shared "room" — a space where human teammates and AI agents operate with the same context, the same resources, and the same conversation history. Instead of an agent living in isolation and reporting back through a narrow chat interface, it becomes a participant in the channel alongside everyone else, seeing what the team sees and responding in the same thread.
Crucially, Switch is framework-agnostic. It's built to coordinate agents built on Claude Code, LangChain, Google's ADK, OpenAI's tools, AWS Bedrock, and custom in-house agents — all inside the same room, without requiring a team to migrate everything onto a single agent framework or vendor. It also isn't limited to the big three chat platforms: SandboxAQ says it integrates with Telegram and Mattermost as well. The project is open-source and self-hostable, which matters for any team that needs to keep agent traffic and conversation data inside their own infrastructure rather than routing it through a third-party SaaS layer.
Plenty of tools let you bolt an AI chatbot onto Slack. What makes Switch notable is the framing: it's explicitly positioned as connective infrastructure rather than a replacement product. SandboxAQ's own description is direct about this — Switch "does not replace the stack you already have. It connects it." For a team that has already invested in Claude Code for engineering, a LangChain agent for support triage, and a custom Bedrock model for internal reporting, the alternative to Switch is usually three different integrations, three different notification patterns, and three different places for a human to check. Switch's pitch is that all three can show up in the same channel, coordinated through shared context, without a rebuild.
That "no migration, no lock-in" design is also why it's drawing attention from developers who are normally skeptical of yet another orchestration layer — the open-source, self-hostable nature means adopting it doesn't mean betting your agent infrastructure on a single vendor's roadmap.
This is squarely relevant to any team already running — or about to run — multiple AI agents across different frameworks and wanting them visible to the humans who need to act on their output. Automation and no-code shops building client-facing agent workflows, internal ops teams stitching together agents from different vendors, and dev teams that have standardized on tools like Claude Code but still need updates to land where the rest of the company already works, are the obvious fits. If your team's agents currently live in a tangle of separate dashboards and Slack webhooks that nobody fully trusts, Switch is worth a look specifically because it's free, open-source, and doesn't ask you to abandon what you've already built.
Switch is available now on GitHub under SandboxAQ's sandbox-quantum/switch repository, free to self-host. Because it's designed to sit alongside existing agent frameworks rather than replace them, the practical first step for most teams is identifying which agent already produces the most "someone needs to see this now" output — a monitoring agent, a support triage bot, a build/deploy notifier — and routing that one into a shared Slack or Teams channel first, before expanding to a full multi-agent room.
RAG assistants on your documents, AI agents that act in your tools, and LLM features inside your product.
Our cutting-edge features simplify collaboration and creativity, making your workflow intuitive and efficient. Transform your vision into reality effortlessly with Hadidiz Flow.
