TypeScope introduces Jo, a statically typed programming language that enforces fine-grained security permissions through the type system for untrusted code execution.

TypeScope has introduced Jo, a programming language where side effects are denied by default and authority must be granted explicitly through capabilities checked by the compiler. The language is designed to restrict untrusted programs—such as AI-generated code—to only the fine-grained capabilities they have been granted, at the level of precision real systems need: a specific directory, a single API host, a read-only interface, or only database rows belonging to the current user.

Jo uses capability-based programming, where authority is represented by explicit parameters that can be passed, refined, substituted, and restricted. The compiler tracks which capabilities code may use, so confinement is expressed in interfaces and types rather than hidden in runtime configuration, combining object-oriented and functional programming with type inference, pattern matching, and context parameters.

Jo is early-stage software with an extensive test suite and a core capability model ready for experimentation. The project is open source under the Apache License 2.0 and welcomes feedback from language designers, security engineers, and developers building agentic systems; the repository is available at github.com/typescope/jo.