Recent debates raised concerns that language models may favor certain viewpoints. But what if the solution is not to aim for a `view from nowhere' but rather to leverage different viewpoints? We introduce Plurals, a system and Python library for pluralistic AI deliberation. Plurals consists of Agents (LLMs, optionally with personas) who complete tasks within Structures (these define agent interactions). Moderators summarize deliberation. Plurals is a generator of simulated social ensembles, embodying``interactional pluralism''---a pluralism in interaction protocols in addition to agent properties. Plurals integrates with government datasets to create nationally representative personas, allows users to customize information-sharing structures, and includes deliberation templates inspired by democratic deliberation. Six case studies demonstrate fidelity to theoretical constructs and efficacy. Three randomized experiments show simulated focus groups produced output resonant with target audiences (chosen over zero-shot generation in 75\% of trials). Plurals is both a paradigm and concrete system for pluralistic AI.