Devotd Github
NewAI agents

arjunvlama/toli v0.1.2 — A collaboration daemon for agents across environments

arjunvlama/toli is a collaboration service that lets agents running on different tools or different computers hand work back and forth through a daemon.

Repositoryarjunvlama/toli ↗

arjunvlama/toli is a collaboration service that lets agents running on different tools or different computers hand work back and forth through a daemon.

GitHub description: Let Claude Code, Codex, OpenClaw or any other agent work together, on one machine or across many.

stars
★ 0
Forks
2
Language
Go
License
Apache-2.0
Latest release
v0.1.2
Read original source ↗View on GitHub ↗

arjunvlama/toli is a collaboration service that helps agents including Claude Code, Codex, and OpenClaw work together on one computer or across several computers. The latest release is v0.1.2. The provided information does not list what changed in this version, so the specific updates cannot be confirmed. It is written in Go and uses the Apache-2.0 license.

Each agent connects to the toli daemon as an MCP client, and the daemons use A2A to exchange task delegations, questions, and responses. In other words, instead of agents connecting directly to one another, the daemons take care of passing work along. A human registers which peer agents are allowed. Connected agents exchange work, and the work can be inspected and managed from a terminal or command-line tool. The core of this project is structuring collaboration between agents in different tool environments and on different computers as daemon-to-daemon communication.

Even when it is used across multiple computers, directly exposing the communication address to the internet is not recommended. The README advises using an encrypted tunnel or a reverse proxy for external network connections. Therefore, the network connection path has to be configured securely, separately from the daemon-to-daemon collaboration feature. The provided information explains this security boundary, but does not present specific changes in the release or performance figures.