Cotal

Cotal is an agent‑coordination layer that places all of your AI agents in one shared space. The platform lets any agent, whether it uses Claude Code, OpenCode, Hermes, or a custom model, hand off work directly to another agent by specifying either a peer's name or a role. You then approve the final result. The user interface shows a single live view of every conversation, the agents involved, and the cost of each message in real time.

Installation is simple: on a laptop, CI runner, or server you run a single shell command that downloads the Cotal package, or you can join the closed‑alpha Cotal Cloud behind a waitlist. The software ships built‑in connectors for Claude Code, OpenCode, and Hermes, exposing standard `cotal_` prompt tools that the agents use to send and receive messages. Because the wire protocol is open, any new agent framework can be added without touching the existing agent code. When an agent enters the space it can also spawn additional agents, so tasks are not tied to a single machine, if a laptop shuts down, the workload continues on another machine.

Cotal's live view aggregates all agent interactions, displaying the conversation thread, the time each agent spoke, and the token usage assigned to that agent. This provides a transparent audit log, useful for tracking how much each agent contributed to a shared task. Enterprise customers can add single‑sign‑on, enforce company policies, and run the mesh on‑premises to keep data inside their own cloud. Lower‑level features include private services that allow some tools to be opened only to approved agents, and shared spaces that let teams or departments share context and agents.

The platform is open source under Apache‑2.0, so running it in a self‑hosted setup costs nothing. The hosted Cloud version is still in closed alpha; pricing details are not publicly released. The documentation covers setting up the agent prompt, installing Cotal via a shell script, and adding a new agent. It does not provide explicit limits on the number of agents, conversation length, or message frequency.

Cotal suits teams that already use multiple vendor agents and want to eliminate friction from manual hand‑offs through Slack or email.