AgentConnect today introduced an open-source platform designed to enable teams and multiple AI agents to collaborate across the communication and development tools they already use. The announcement addresses a growing challenge in the AI industry: while individual agents have become increasingly capable, their work often remains isolated to single-user terminals, leading to fragmented context and duplicated efforts.
According to the company, most agent work begins and ends with one person, with sessions confined to individual terminals and context siloed. Teams frequently build separate bots and integrations for each new workflow, and existing solutions that add a team layer often force users into dedicated workspaces or tie them to a single provider. AgentConnect aims to change this by bringing agents directly into existing team conversations, such as Slack threads, Telegram chats, and Discord channels.
The platform supports ACP-compatible agents like Claude Code and Codex, allowing teams to assign distinct roles to each agent. This enables human and AI participants to collaborate through shared conversations and agent-to-agent calls. For instance, during an incident response, a triage agent can inspect context, a code specialist can trace a regression, and a verification agent can check proposed results—all while human teammates provide input and follow the process in the same thread.
The implications extend beyond incident response. Customer support issues can move from a public Telegram conversation to a trusted Slack workspace without manual context transfer. Engineering teams can combine general and specialist reviewers for pull requests, with inline GitHub comments and independent check results. Scheduled and webhook-driven tasks can also pause for human input when needed.
Key features of AgentConnect include support for multiple agents with distinct configurations, integration with existing tools such as Slack, Telegram, and Discord, and fine-grained access control. Administrators can manage agent access separately from session visibility, control which tools and repositories agents can reach, and link multiple social accounts to a single member profile for consistent permissions.
The platform is open source under the Apache 2.0 license and can be self-hosted via Docker Compose or Kubernetes. This ensures that agent execution and workspaces remain within the team's controlled environment. The announcement highlights a growing trend toward open-source, interoperable AI tools that prioritize flexibility and data sovereignty.
AgentConnect OSS is available today. The project is hosted on GitHub, with additional information at agentconnect.md and full documentation available online.
