Peer reviewer finder and invitation desk for journal editors

I edit an academic journal and I want an app I can work out of to find qualified peer reviewers for each submission, instead of hunting through search engines by hand every time a manuscript lands. Build me an editor-side reviewer sourcing desk on top of Crossref, Google Sheets, and Gmail.

The app opens on a manuscript queue. Read it from a Google Sheets tab with Get Values, where each row is a submission with title, abstract, author names, and status. Show the queue as a list I can scan, and clicking a manuscript opens its reviewer board.

The reviewer board lists ranked candidate reviewers for that manuscript. For each candidate show their most relevant recent papers, citation counts, when they were last invited, and a conflict flag. Anyone already invited in the last 90 days is greyed out as fatigued but still visible, so I can see who has been asked recently and why they are unavailable. Candidates whose conflict flag is raised stay on the board with the reason shown rather than being hidden, because I want to clear or drop them myself.

Each manuscript has a Find reviewers button that kicks off a background agent. The agent searches Crossref with Search Works, using the manuscript title and abstract as a bibliographic query (query.bibliographic), filtered to journal articles from the last five years (filter type:journal-article and from-pub-date), sorted by relevance score or is-referenced-by-count descending. It pulls the author lists off the top matches with Get Work by DOI, then ranks recurring authors by topical overlap across those matches and by citation weight, so someone who appears on several close papers with well cited work rises to the top.

The same agent then screens each candidate for conflicts. For every shortlisted person it searches that individual's own publication history with Search Works using the field scoped query.author parameter over the last five years, and flags anyone who co-authored with a manuscript author in that window or who lists the same affiliation. Calibrate this to favor recall over precision: it is better to flag a potential conflict I can evaluate and clear than to miss a real one. Write the flag with its reason attached, for example recent co-authorship on a named paper, or shared institution, so I can judge it at a glance.

Two rules are non negotiable in the shortlist. Always exclude the manuscript's own authors and their recent co-authors from the ranked candidates entirely. And where Crossref has no affiliation data for a person, show it as unknown rather than scoring it as conflict free, since affiliation is deposited inconsistently and an empty field means missing data, not absence of a connection. Co-authorship overlap is the reliable signal and affiliation matching is best effort.

Each candidate also has a Draft invite button. It uses Gmail Create a Draft to stage a personalized invitation that names the candidate's most relevant paper and connects it to the manuscript topic, addressed to them and left in my drafts so I can review and send it myself. Nothing sends automatically.

Invited, accepted, and declined statuses write back to the sheet using Append Values for new reviewer activity rows and Update Values for status changes on existing rows, including the invitation date. The reviewer pool view reads that history back so the 90 day fatigue greying stays accurate across sessions and across manuscripts.

One Crossref implementation detail: include a mailto query parameter on every request so the app sits in the polite pool and does not get throttled when the agent fans out across many candidate publication history lookups. Handle the response envelope where list results live under message.items and single work records live directly under message.