Foundations
A suite of tools for scaling software development at the pace of AI
A way to model users, stakeholders, internal staff, and anyone engaging with what you build. Personas for all the different edge cases, competitors, power users, and bad actors the software will be in contact with.
By embodying these different cohorts and perspectives the feature set, code, and UX can be tested, scored, and judged to find gaps, explore new ideas, and catch security issues.
Learn more →A full catalog of UI components that act as a shared language that agents can read. They can be easily referenced, imported, or copied, and each comes with full behavior specs, props, state, design tokens, and edge cases.
Instead of building with a library or from an LLM's best guess, the agent can get exact replicable instructions to build a polished experience.
Learn more →Empowers agents to automatically improve the quality of the software on the fly.
By tracking the right metrics agents can be signaled to fix issues and improve code quality in the background. UX and code quality improve purely from usage.
A stack of tools to drive work to completion. By creating audits, tests, scoring, and criteria a project, feature, or system can be fully fleshed out with minimal human intervention.
Multiple agents can work on a problem together, grading each other's work, and scaling as much as their scoping allows. Allows orchestrated systems to get large amounts of work done in the background.
A store for the information pollution that can seep into agentic systems. Confidential information, names, overly specific instructions, and unneeded emphasis all start to pollute the code, docs, text, and even the production UI and copy.
By capturing and cataloging agentic smell it can be cleaned out of the system.
Join the waitlist
Foundations is invite-only while we build it. Add your email and we will let you know when spots open.