A small company with a long view.

RBJ Global is a portfolio of independent, privacy-first software, built and operated by a single principal.

Each product is its own application, on its own domain, with its own roadmap. Today that includes a desktop AI workspace, a free AI-education library, a local-first dictation tool, a multi-agent stock research framework, a browser-based document converter, and an open-source coordination server for AI agents. Every one is live, in real use, and built to production standards: a full marketing site, end-to-end test coverage, a deployment pipeline, and documentation.

What ties them together is not a large team. There isn't one. The portfolio is produced by one principal directing a coordinated set of AI agents, built on the same kind of tooling published openly under this portfolio, including ClaudeLink, the open-source layer that lets those agents work as a team. The work a conventional studio distributes across product managers, designers, engineers, and QA is instead held by one person who owns both the domain knowledge and the orchestration. The work is checked through automated test suites rather than handed between roles.

The result is a different cadence. Software the older model measured in quarters, and staffed with cross-functional teams, can be designed, built, tested, and shipped by one operator in a fraction of the time, without lowering the bar on testing, documentation, or polish. The portfolio is the evidence; it exists so the claim can be checked rather than asserted.

The engineering behind each product, including the architectural decisions, the failures, and what was learned from them, is documented in the case studies at rbjglobal.com/engineering. They are written in the same plain, factual register as the software.

The portfolio continues to grow.