Tetrees Agent is a runtime platform for discovering, running, and growing compact agent assets above the model layer. It supports multiple large language models, including OpenAI, Claude, and Z.AI, via a unified web interface and Model Context Protocol runtime. Using the TAIP/1 format, the platform standardizes agent intelligence, memory policies, tools, and evaluations across providers. The Agent AVCP audition feature ensures structural integrity, safety compliance, model compatibility, regression evidence, and ownership binding. Users can execute models with Tetrees-hosted options using a Point balance system or their own provider keys, which are never stored. Growth mechanisms let owners review and accept useful memories, creating checkpoints with rollback options. Integrated with MCP 2.2, Tetrees Agent allows users to manage AI Packs through a single revocable account token, supporting search, acquisition, execution, growth, and publishing operations.
