omniseek_paper_recommend
Recommends semantically-similar papers from a seed paper ID using SPECTER embeddings, surfacing conceptually related work, including recent papers, that keyword search and citations miss.
Instructions
Use WHEN you have a paper and want more like it — semantically-similar papers (SPECTER embeddings) that keyword search and the citation graph miss, including very recent work. Uses Semantic Scholar's recommendation model (SPECTER embeddings + co-citation), so it surfaces conceptually-related work that omniseek_search (keyword) and omniseek_field_skeleton (citations) miss — including very recent papers the citation graph has not caught up to.
Fully-qualified MCP name: mcp__omniseek__omniseek_paper_recommend (server name is omniseek; there is no omniseek-eye server).
Pass seed paper ids (arXiv ids / DOIs / S2 ids — a paper you found via omniseek_search or omniseek_field_skeleton). One seed = "more like this"; several = recommendations from that set. This is OmniSeek's "semantic search": it routes to S2's existing embeddings rather than building any. For an openalex omniseek_search result pass metadata.paper_id (or metadata.doi), NOT source_id — the OpenAlex W-id is a graph id the paper tools do not accept.
Returns: {"seeds", "n", "papers": [{id, title, year, date, cited_by, first_author, doi, url}]} (ordered by S2 relevance; YOU re-judge). Citation neighborhood instead → omniseek_field_skeleton; keyword search → omniseek_search.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| ids | Yes | ||
| limit | No |