Switch Brings Open-Source AI Teammates Into Existing Chat Platforms

Overview of Switch

Switch is an open-source project designed to integrate AI agents directly into popular team communication platforms such as Slack, Microsoft Teams, Discord, Telegram, and Mattermost. Rather than requiring users to adopt a separate application or chat interface, the platform turns existing communication channels into collaborative workspaces where humans and AI agents work side by side.

According to the project documentation, Switch acts as a connective layer rather than a standalone chat app or an AI model provider. Organizations can self-host the software on their own infrastructure, maintaining control over their data, context, and configurations while using external agent harnesses and language models of their choice.

Key Architectural Features

• Multi-Platform Integration: Works within Slack, Teams, Discord, Telegram, and Mattermost channels.
• Model and Harness Agnostic: Supports various agent frameworks (such as LangChain) and coding assistants (such as Claude Code or Codex) as long as they adhere to supported protocols.
• Role-Based Access: Agents can be assigned specific roles, instructions, and tool integrations, including access to services like Google Workspace, Jira, and GitHub.
• Self-Hosted Context: Because the infrastructure runs locally or on private servers, team context remains internal, allowing organizations to switch vendors or models without losing historical data.

Practical Workflows and Use Cases

The software aims to move beyond single-turn question-and-answer bots by enabling continuous, multi-step workflows. Reported use cases include:

• Incident Management: An agent can detect error spikes, open an incident channel, pull relevant logs, and coordinate diagnostics.
• Triage and Support: Incoming bug reports or feature requests can be automatically analyzed, reproduced where possible, and filed into tracking tools like Jira, Confluence, or Notion.
• Collaborative Coding: Developers can work alongside coding agents in dedicated feature channels, bringing in human colleagues for reviews without needing to re-explain context.

Verification and Telemetry

The project provides transparent documentation and allows users to inspect its codebase directly on GitHub. Switch Console, the desktop companion application, includes telemetry features to collect anonymous usage data for improvement purposes. According to the disclosures, telemetry is opt-out in the desktop client and off by default in the server core. The documentation emphasizes that usage analytics are restricted to fixed enumerated categories, numerical counts, and durations, explicitly excluding source code, prompts, personal data, and free-text fields.

To verify official announcements, review source code, or check deployment instructions, users can consult the repository directly at the official GitHub project page or review the Model Context Protocol (MCP) documentation hosted by the project.