Hey Product Hunt! We're the makers of Agentero — a local-first research workbench built for the agent era. The problem we kept hitting: research context is fragmented. PDFs and highlights live in Zotero, notes in Obsidian, and your discussions with AI stay trapped in chat windows. None of them talk to each other, so even the smartest agent only ever sees fragments. And as models get smarter and more alike, what sets output apart is context — your context. So we built Agentero around one idea: keep your whole research context in a local Vault of plain Markdown files, and let any agent you choose work with it. What you can do today: - Bring your own agent via ACP — no lock-in to any model or vendor — to summarize, question, and translate with full document context - Migrate from Zotero in one click, keeping tags, notes, and attachments - Import papers from DOI/arXiv, RSS feeds, or the Zotero browser extension - Read split-screen: PDF on the left, agent on the right, with highlights, annotations, and translation - Write notes with Obsidian-style [[wikilinks]] and export BibTeX/BibLaTeX for LaTeX - Plug external tools like ChatGPT straight into your library via the built-in CLI and MCP Agentero is free and open source (MIT) on macOS, Windows, and Linux. We'd love your feedback: what would you want your agent to do with your paper library?

Agentero

Agentero — Agent friendly research workbench

Agentero — Agent friendly research workbench