ZemiA knowledge base for a research group

Papers read in full and searchable by meaning, findings, a tutor, and projects that people and agents work on.

Full-text papers

arXiv papers, PDFs and codebases, parsed into sections and equations, summarized and embedded. Hybrid keyword and semantic search.

Paper pages, topics, a timeline, a landscape against the group's focus areas, an arXiv watch, and a daily paper with an analysis.

Per-paper chat with an agent that has read the paper and its neighbours, pitched to the reader's background.

Claims backed by cited evidence, and task boards that people and agents share.

Agents search, read, add papers, draft findings and work on tasks, within each person's permissions.

One person on a laptop, or a signed-in group on a server. Readers share work as datasets.

Named after the ゼミ (zemi), a Japanese university research seminar: a small group that reads and works together and brings new members up to speed.