White papers
Long-form papers on the design and engineering decisions behind GO Pilot GO.
2026-04-05 · 14 pages
Voice-first agent UX: design principles for sub-three-sentence agent creation
A design study on the constraints, affordances, and failure modes of voice-first agent builders, with reference patterns from GO Pilot GO.
Read paper →
2026-04-12 · 18 pages
Programmatic sub-agent platforms: a primer for AI engineers and planners
Why the next layer of the AI stack is not bigger models but a hosted catalog of cheap, callable specialist agents that planning models can invoke.
Read paper →
2026-05-10 · 12 pages
Resource metering for AI agents: a model for tokens, voice, and browser seconds
Per-run pricing is a relic of pre-agentic tooling. This paper proposes a three-axis meter (tokens, voice seconds, browser seconds) and a workspace daily cost ceiling as the right pricing primitive for agent platforms.
Read paper →
2026-05-25 · 16 pages
Per-user encrypted OAuth architecture for multi-tenant agent platforms
A reference architecture for storing, scoping, and rotating end-user OAuth credentials in a platform where agents act on behalf of individual operators.
Read paper →
2026-06-01 · 15 pages
An evaluation rubric for picking an agent platform in 2026
Twelve criteria for evaluating agent platforms, weighted by what actually matters in production: cost predictability, credential safety, callability, and operator authoring speed.
Read paper →