← Files Paper NexusARCHIVED FILE
plugin.json
8.28 KB · Nov 18, 2026 · 06:19 UTC
{
"$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
"name": "papernexus",
"version": "1.0.0",
"description": "Find related scientific papers through their citation network: search the scholarly literature, get recommendations from one or more papers, read a paper's references and citing papers, and browse your Paper Nexus library.",
"author": {
"name": "Paper Nexus",
"email": "info@papernex.us",
"url": "https://papernex.us"
},
"homepage": "https://papernex.us",
"repository": "https://github.com/PaperNexus/papernexus-mcp",
"license": "MIT",
"keywords": [
"research",
"papers",
"citations",
"literature-review",
"academic",
"science",
"openalex",
"semantic-scholar"
],
"extensions": {
"com.openai": {
"interface": {
"displayName": "Paper Nexus",
"shortDescription": "Find related scientific papers",
"longDescription": "Paper Nexus maps the scientific literature by its citations, using the scholarly works in OpenAlex and Semantic Scholar. Find a paper by its title, a DOI or a link; get the papers related to one paper or several through the citation network, each with the citation links that put it there; read a paper's details and most-cited references; see the papers that cite it; narrow recommendations and citing papers with filters (years, citations, authors, venues, title words, citing or cited papers, open access); and browse your Paper Nexus library, if you allow it. Every paper links to papernex.us, and every set of recommendations comes with a link to the whole citation graph.\n\nIt is for researchers, students and librarians who are starting a literature review, checking references, or looking for what to read next.\n\nLimitations: Paper Nexus holds papers' metadata and the citations between them, not their full text. A paper with few or no citation links gets few or no related papers. Each account has a limit on recommendation and citing-paper lookups, and get_usage shows what is left.\n\nSign in with your Paper Nexus account, or create one when you connect.",
"developerName": "Paper Nexus",
"category": "Education & Research",
"capabilities": [
"Search papers by title, DOI or link",
"Recommend papers related to one or more papers",
"Read a paper's details and most-cited references",
"List the papers citing a paper",
"Filter recommendations and citing papers",
"Read your Paper Nexus library"
],
"websiteURL": "https://papernex.us",
"supportURL": "https://papernex.us/contact",
"privacyPolicyURL": "https://papernex.us/privacy",
"termsOfServiceURL": "https://papernex.us/terms",
"defaultPrompt": [
"Find papers related to \"U-Net: Convolutional Networks for Biomedical Image Segmentation\".",
"What should I read next after the ResNet paper (arXiv 1512.03385)?",
"Show me the main references of the AlphaGo paper (\"Mastering the game of Go with deep neural networks and tree search\")."
],
"brandColor": "#5566EE",
"composerIcon": "./assets/logo.png",
"logo": "./assets/logo.png"
},
"review": {
"test_cases": {
"positive": [
{
"description": "Related papers for one paper, found by its title.",
"prompt": "Find papers related to U-Net: Convolutional Networks for Biomedical Image Segmentation.",
"tools_triggered": "search_papers, recommend_papers",
"expected_behavior": "search_papers finds \"U-Net: Convolutional Networks for Biomedical Image Segmentation\" (2015). recommend_papers with its id returns related papers, most related first, mostly image segmentation work such as \"Fully convolutional networks for semantic segmentation\", \"SegNet\" and \"V-Net\", along with medical image segmentation papers. Each paper has a papernex.us link, and the answer links to the whole citation graph."
},
{
"description": "A paper's authors, citation count and main references, found from a DOI link.",
"prompt": "Who wrote the paper at https://doi.org/10.1038/nature16961, how often is it cited, and what are its most cited references?",
"tools_triggered": "search_papers, get_paper",
"expected_behavior": "search_papers with the link finds \"Mastering the game of Go with deep neural networks and tree search\" (Nature, 2016). get_paper returns its authors (David Silver first), its citation count (about 19,000) and its most-cited references, among them \"ImageNet classification with deep convolutional neural networks\", \"Reinforcement Learning: An Introduction\" and \"Human-level control through deep reinforcement learning\"."
},
{
"description": "Related papers for two papers at once.",
"prompt": "Recommend papers related to both \"Deep Residual Learning for Image Recognition\" and \"ImageNet classification with deep convolutional neural networks\".",
"tools_triggered": "search_papers, recommend_papers",
"expected_behavior": "Two search_papers calls find the two papers (2015 and 2012). One recommend_papers call with both ids returns convolutional network papers related to them, such as \"Going deeper with convolutions\", \"Very Deep Convolutional Networks for Large-Scale Image Recognition\", \"Visualizing and Understanding Convolutional Networks\" and \"Batch Normalization\", each with a papernex.us link."
},
{
"description": "A title search answered from the search results alone.",
"prompt": "Search for papers titled 'protein structure prediction' and tell me which is most cited.",
"tools_triggered": "search_papers (get_paper optional)",
"expected_behavior": "search_papers returns several papers with that title or one close to it, each with an id, year and citation count. The assistant names the most cited of them (at the time of writing \"Protein Structure Prediction Using Rosetta\", 2004). It may call get_paper on the top results to check their details."
},
{
"description": "A set number of recommendations from a DOI.",
"prompt": "Give me 5 papers to read around doi:10.1038/nature14539, with links.",
"tools_triggered": "search_papers, recommend_papers",
"expected_behavior": "search_papers with the DOI finds \"Deep Learning\" by LeCun, Bengio and Hinton (Nature, 2015). recommend_papers with limit 5 returns five papers, such as \"Learning Deep Architectures for AI\" and \"A Fast Learning Algorithm for Deep Belief Nets\", each with a papernex.us link."
}
],
"negative": [
{
"description": "Nothing to do with scientific papers, so no Paper Nexus tool should be called.",
"prompt": "What's the weather in Paris tomorrow?"
},
{
"description": "A restaurant booking. Paper Nexus finds and reads scientific papers and can't make reservations, so no Paper Nexus tool should be called.",
"prompt": "Book me a table for two at an Italian restaurant tomorrow at 8 pm."
},
{
"description": "Paper Nexus has papers' metadata and citations, not their full text. The assistant may find the paper and link to it, but must say it cannot provide the full text and must not present tool output as the paper's text.",
"prompt": "Give me the full text of the AlphaFold 2 paper."
}
]
},
"commerce": false,
"commerce_description": "Paper Nexus does not sell anything or take payments in ChatGPT."
},
"publication": {
"countries": [],
"release_notes": "First release: search papers by title, DOI or link; recommendations from one or more papers through the citation network, each with the citation links behind it; a paper's details and most-cited references; the papers citing a paper; filters on recommendations and citing papers; the account's remaining lookups; and read-only access to your Paper Nexus library. Every tool that lists papers can return only the fields asked for."
},
"apps": "./.app.json"
}
}
}
SHA-256: 707c6c1c3d8e119ab775607c2d22e6d7409f61a393f3a6ffe0252e5f1263c7e1