Agentero poco-ai

Traditional reference managers are not agent-friendly: highlights and notes stay locked inside single files, making it hard for agents to reuse knowledge across papers; every conversation starts from scratch without a stable local knowledge map; PDFs are human-friendly but not the most comfortable reading material for agents. Agentero aims to build an agent-friendly, agent-native way to manage references, exploring how humans and agents collaborate on reference management. Connect your local agent via ACP (BYOA), select-text chat, import papers and Skills, stay compatible with the Zotero ecosystem, and get paper translation, bidirectional links with a knowledge graph, and deep PDF reading.