Cervellaswarm is an open-source toolkit for enforcing and verifying communication protocols among AI agents using multiparty session types. Its core project, Lingua Universale, allows developers to define, compile, and mathematically verify agent communication rules, preventing protocol violations and deadlocks at runtime. Designed primarily for AI engineers and teams building multi-agent systems, it provides a compiler, standard library protocols, a linter and formatter, LSP server, VS Code extension, REPL, browser-based playground, and extensive CI integration. Cervellaswarm aims to bring the safety and reliability of static typing to autonomous agent conversations, much like TypeScript does for JavaScript.
Visit Cervellaswarm's official website for product details and getting started.
Comprehensive guides and API references for using Cervellaswarm effectively.