An AI interface connects users to library collections

Benefits

Benefit: Northeastern University Library developed a server and Claude plugin that direct academic requests to library resources and return links to underlying works. Library dean Dan Cohen demonstrated the interface in October 2025. A later account described adding museum collections and retrieving relevant material without ordinary web search. Cohen and Evan Simpson subsequently reported an improved discovery experience in early testing. These are accounts from the deploying institution, without a controlled retrieval benchmark or measured learning outcome. [1] [2] [3]
Category: Improved Knowledge Discovery
Status: reported [1] [2] [3]
Beneficiary groups: Organizations.
Expected?: Yes
Outcome identifiers: Not assessed
Beneficiaries: Not assessed
Reach: Not assessed
Distribution and access: Not assessed
Duration: Not assessed
Costs and harms: Not assessed

System: Northeastern University Library MCP server and Claude plugin [1] [2] [3]
Application developer: Northeastern University Library; Anthropic
Workflow: The university connected Claude with its catalog and article resources through a server and direct source links.
Relation to AI: Claude interpreted requests and used a library server to retrieve materials. Cataloging, collections, access controls, and source links were essential parts of the workflow.
AI contribution: Claude interpreted requests and used a library server to retrieve materials. Cataloging, collections, access controls, and source links were essential parts of the workflow. [1] [2] [3]
Model: Not assessed
Release: Not assessed
Role: Not assessed
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Human oversight: Not assessed

Participants: Developer: Northeastern University Library; Anthropic. Operator: Northeastern University Library.
Jurisdictions: Benefit: Boston, Massachusetts, United States. Developer: United States. Operator: United States.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Conditions for replication: The reviewed sources do not establish causal conditions or an effect size for reproducing the benefit.
Replication attempts: Not assessed
Later findings: Not assessed
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Source numbers: Recorded value
Citations: Recorded value [1] [2] [3]
Source relationships: Recorded value [1] [2] [3]
Original evidence: The source review separates original evidence from repeated accounts and records material access limits. [1] [2] [3]
Access and versions: Recorded value [1] [2] [3]
Evidence links: Recorded value [1] [2] [3]
Evidence strength: Recorded value
Language: Not assessed
Supported findings: Not assessed
Missing evidence: Not assessed

Assessed benefits: Organizations.
Reasons: Organizations: The reviewed sources do not establish the magnitude of the benefit.
Observed or intended: Reported benefits.
Fill: The deployer's account
Depth: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed

Field: Education and research. Discovery of scholarly and cultural collections.
Purpose: The library built an AI interface to help users find scholarly and cultural materials while preserving access to source works. [1] [2] [3]
Dates: Recorded value
Stage: Launched library interface with early use and qualitative testing. [1] [2] [3]
Version: 658d2a31eab3b0c162a3b3c55b6d9e4c37c97fced71b6fb6d52a32cf3080136b
Kind: Not assessed
Related entries: Not assessed