SandboxAQ Open-Sources Switch to Connect AI Agents in Slack, Teams
The platform-agnostic tool lets organizations deploy AI agents directly into existing collaboration channels at no cost.
SandboxAQ has released Switch, an open-source platform that allows organizations to deploy AI agents directly into the collaboration tools their teams already use, including Slack, Microsoft Teams, and Discord.
The software addresses a common enterprise challenge: AI agents built with different frameworks often operate in isolation, requiring repeated context sharing and manual coordination between tools. Switch transforms standard collaboration channels into shared workspaces where human team members and AI agents work side-by-side with access to the same conversation history, context, and resources.
Platform-agnostic architecture
Switch is designed to work with agents built on any major framework, including Claude Code, Google ADK, LangChain, and OpenAI. The platform connects through MCP (Model Context Protocol), APIs, and adapters, allowing teams to use agents from multiple providers without locking their workflows into a single AI vendor's ecosystem.
The software can be deployed in minutes within standard enterprise environments. SandboxAQ offers Quick Start and Team editions, with an extensible architecture for enterprise-scale deployment. The platform is released under Apache 2.0 with the Commons Clause and is free for commercial and personal use.
Persistent context across workflows
A key feature of Switch is its room-based architecture. Each channel or room maintains its own context, participant list, and rules as team members and agents join or leave. This means specialized agents can be added to ongoing projects—such as live incident response or campaign development—without requiring manual briefings, because the room already contains the full timeline and relevant information.
A single agent can operate across multiple rooms simultaneously, with each maintaining separate context. According to SandboxAQ, this design allows teams to reuse agents across projects while preserving the specific history and rules of each workspace.
Why it matters
As enterprises adopt multiple AI agents for different functions, coordination overhead threatens to erase productivity gains. Switch tackles this by embedding agents in the collaboration platforms where work already happens, rather than forcing teams to context-switch between separate AI tools. The open-source, vendor-neutral approach also reduces the risk of AI platform lock-in—a growing concern as organizations build agent-dependent workflows.
Part of broader agent portfolio
Switch is part of Flint AI, SandboxAQ's portfolio of AI agent products. The company positions it as a natural progression from Flint CLI, a free tool developers use to test and evaluate individual agents locally, to production deployment where multiple agents collaborate with human teams.
Mohammed Aboul-Magd, General Manager at SandboxAQ, said the company built Switch using Switch itself, with engineering, design, and marketing teams collaborating inside rooms throughout development and launch.
The software is available now on GitHub with tutorials and documentation at FlintAI.dev, according to details first reported by SandboxAQ in its announcement.
This is an original analysis by the Omega editorial team. Source reporting: AI Watch.
Want systems like this working for your business?
Book a Call