LibCorpus API

Version 1.0.0 · OpenAPI 3.1 document at /openapi.json

LibCorpus serves the figure data of open-access papers in quantum science and technology. Every plotted panel is a CSV with its columns, provenance and licence; every paper and panel has a stable page, a Markdown digest and a JSON record.

No key, no registration. Requests from any origin are answered (CORS). The machine-facing routes share a ceiling of 600 requests a minute per client; beyond it the server answers 429 with a Retry-After header.

Stability: the v1 URL shapes and the fields of their responses do not change. Fields may be added. Missing values are null. Lists come in pages: {total, count, offset, next, items}, where next is the following page's URL. Errors are JSON, {error: {code, message}}, with code one of not_found, invalid_id, invalid_parameter, rate_limited or internal; an unknown query parameter is an error rather than ignored.

Data stays under the licence or terms shown on each paper: license.data in the records, and the header of every CSV. Cite the original paper; the citation files name the LibCorpus page and package version in their note.

A plain-text guide for language models is at https://libcorpus.org/llms.txt. The same corpus is available as a tool for AI assistants over the Model Context Protocol at https://libcorpus.org/mcp (Streamable HTTP, stateless, no authentication): add that address as a remote MCP server in Claude, ChatGPT, Cursor or any MCP client and it gains search_papers, list_papers, get_paper, get_panel, get_panel_data and cite.

Start here

  1. Find the paper: GET /api/v1/search?q=… with words, a DOI or an arXiv id, or list everything with GET /api/v1/papers.
  2. Read it: GET /api/v1/papers/{paperId} gives every figure and panel with its description and links.
  3. Plot a panel: GET /api/v1/papers/{paperId}/figures/{fig}/panels/{panel} for the columns, their roles and each file's layout, then the file's url for the values.
# pip install pandas matplotlib requests
import matplotlib.pyplot as plt
import pandas as pd
import requests

panel = "https://libcorpus.org/api/v1/papers/LCP000034/figures/fig1/panels/e"
# The panel record names each column's role (x, y, ...), symbol and file.
rec = requests.get(panel).json()
cols = {c["role"]: c for c in rec["columns"]}   # this panel: one x, one y
x, y = cols["x"], cols["y"]
# The lines starting with # are the provenance header.
df = pd.read_csv(f"{panel}/files/{y['file']}", comment="#")
ax = df.plot(x=x["name"], y=y["name"], legend=False)
ax.set(xlabel=x["symbol"], ylabel=y["symbol"])
plt.show()
print(rec["citation"]["text"])

It draws Fig. 1(e) of LCP000034, a fluorescence dip at zero field, from the columns the panel record names for x and y; compare it with the panel on the site. The data are CC BY 4.0, and the last line prints the citation. A panel with several curves says which columns form each one under series; a heatmap adds a z column.

Prefer to hand a model the whole thing as text? Every paper is at /paper/{paperId}.md, and /llms.txt explains the library in one page.

Connect an AI assistant

LibCorpus is also an MCP server. MCP (Model Context Protocol) is the standard way to give an AI assistant tools. Add the address below to Claude, ChatGPT, Cursor or another MCP client, and the assistant can find a paper by its DOI, see which figures have data, read the values behind a panel, plot them and cite the paper, all from the conversation. It needs no account or key, and it only reads.

Once it is added, you can ask, for example:

1. Copy the server address.

https://libcorpus.org/mcp

2. Add it to your assistant. Wherever an app asks about sign-in or authentication, choose none.

Claude (claude.ai and Claude Desktop, every plan; Free allows one custom connector)

Customize, then Connectors: + Add, then Add custom connector. Name it LibCorpus, enter the address as the server URL and choose Continue; review the detected settings and Continue again; under Authentication choose No sign in, then Add. In a chat: +, then Connectors, and turn LibCorpus on. On Team and Enterprise an owner adds it first (Organization settings, then Connectors, Add, Custom, Web); members then find it under Customize, Connectors and choose Connect.

Claude Code

Run in a terminal:

claude mcp add --transport http --scope user libcorpus https://libcorpus.org/mcp

Without --scope user it is added to the current project only.

ChatGPT (on the web; not every plan or workspace allows custom MCP servers)

Open Plugins (chatgpt.com/plugins): +, then Add custom MCP server. Name it LibCorpus, enter the address as the Server URL, choose No authentication, confirm the warning, and choose Create as a plugin. Then install the plugin: a plugin that is created but not installed cannot be used. In a new chat, type @ and pick LibCorpus.

Cursor

Add to ~/.cursor/mcp.json, or to .cursor/mcp.json in one project:

{
  "mcpServers": {
    "libcorpus": { "url": "https://libcorpus.org/mcp" }
  }
}

VS Code

Run MCP: Add Server from the Command Palette and choose HTTP, or add to .vscode/mcp.json:

{
  "servers": {
    "libcorpus": { "type": "http", "url": "https://libcorpus.org/mcp" }
  }
}

VS Code needs the "type": "http", which Cursor's form leaves out.

Windsurf (now Devin Desktop)

In the Cascade panel, the ... menu, then Open MCP config file; add this under mcpServers and save:

{
  "mcpServers": {
    "libcorpus": { "serverUrl": "https://libcorpus.org/mcp" }
  }
}

Checked against each app's own documentation on 19 October 2026, not tried in every plan. Apps move their menus, so a label may since have changed.

3. Check that it works. Ask: "Which LibCorpus tools can you use?" It should name six: search_papers, list_papers, get_paper, get_panel, get_panel_data and cite. If it names none, see the list below.

Then: "Using LibCorpus, find the paper with DOI 10.1103/PhysRevApplied.23.034001, describe Fig. 1(e), and plot it from all of its data." The answer should name LCP000034, a near-zero-field fluorescence spectrum of nanodiamonds (normalized fluorescence against B in mT, 80 points), and cite the paper.

If it does not work

No option to add a custom connector or MCP server.
Your plan or workspace may not allow one. On Claude Team or Enterprise, and in a ChatGPT workspace, an owner or admin adds or allows it first. The Claude Free plan allows one custom connector.
Asked to sign in or for a key.
Choose No sign in or No authentication. LibCorpus needs neither.
The assistant answers without using LibCorpus.
Turn it on for this chat (Claude: +, then Connectors; ChatGPT: type @ and pick LibCorpus), start a new chat after adding it, and name LibCorpus in the question.
A search finds nothing.
Search matches every word given, so long sentences rarely match. Ask for a DOI, an arXiv id or two or three keywords. An empty result means no match in LibCorpus, not that no such paper exists.
The answer says the data was shortened.
get_panel_data returns 500 rows unless asked for more, at most 5,000 per answer, and says which rows came back. For anything about the whole panel (extremes, fits, totals), ask the assistant to read every row: it continues with offset until the answer is complete, or loads the CSV from the address it gives, which needs a code tool. You can also download the file yourself.

MCP tools

The assistant chooses the tools itself; you ask in your own words. These six are what it can call. A ? marks an optional argument.

search_papers(query, limit?, offset?)     find papers by id, DOI or keywords
list_papers(limit?, offset?)              every paper, newest first
get_paper(paper_id, figure?, format?)     a paper's figures and panels
get_panel(paper_id, figure, panel)        a panel's columns and files
get_panel_data(paper_id, figure, panel,   the plotted values as CSV
  file?, columns?, offset?, max_rows?)
cite(paper_id, figure?, panel?, format?)  cite as text, BibTeX or CSL-JSON

Every answer is held under about 40,000 characters, so it fits in an assistant's context. Longer results come in pages: search_papers, list_papers and get_panel_data say how many there are in all, give a next_offset to continue from and say whether the answer is complete. A shortened answer says what it left out and gives the address of the whole file. All six tools only read.

search_papers

Finds papers by a LibCorpus id, a DOI, an arXiv id or a few keywords. Every word must match, in identifiers, titles, authors, abstracts or panel descriptions, so short queries work best. Each result names up to three matching panels.

search_papers(query, limit?, offset?)
query requiredAn identifier, or one to three keywords.
limitResults per answer, default 20, at most 100.
offsetResults to skip, default 0. Pass the previous answer's next_offset to continue.

Ask: "Which LibCorpus papers are about Rydberg atoms?" The assistant might call it with:

{"query":"Rydberg"}

list_papers

Lists every paper in the library, newest first, with its id, title, authors, year, DOI, arXiv id and number of panels.

list_papers(limit?, offset?)
limitPapers per answer, default as many as fit, at most 500.
offsetPapers to skip, default 0. Pass the previous answer's next_offset to continue.

Ask: "What has been added to LibCorpus recently?" The assistant might call it with:

{"limit":10}

get_paper

One paper: its metadata, abstract, license and citation, and each figure with its panels and what they plot. A large paper comes back shortened and says so; ask for one figure to get it in full.

get_paper(paper_id, figure?, format?)
paper_id requiredA LibCorpus id, such as LCP000034. search_papers turns a DOI or an arXiv id into one.
figureOne figure only, in full: fig1, extfig2.
formatjson (default) or markdown.

Ask: "Which figures of LCP000034 have data, and what does each panel show?" The assistant might call it with:

{"paper_id":"LCP000034","figure":"fig1"}

get_panel

What it takes to plot one panel: the plot type, each column with its role (x, y, z, error, band) and its symbol as plotted, which columns form each curve, and each file with its address, size and sha256.

get_panel(paper_id, figure, panel)
paper_id requiredA LibCorpus id, such as LCP000034. search_papers turns a DOI or an arXiv id into one.
figure requiredA figure id as get_paper lists it: fig1, fig2, extfig1.
panel requiredA panel id as get_paper lists it: a, b, d-inset.

Ask: "What are the axes and units of Fig. 1(e) in LCP000034?" The assistant might call it with:

{"paper_id":"LCP000034","figure":"fig1","panel":"e"}

get_panel_data

A panel's plotted values as CSV, with the provenance header first and a line saying which rows came back of how many. Rows come in pages: the assistant continues with offset until the answer is complete before saying anything about the whole panel.

get_panel_data(paper_id, figure, panel, file?, columns?, offset?, max_rows?)
paper_id requiredA LibCorpus id, such as LCP000034. search_papers turns a DOI or an arXiv id into one.
figure requiredA figure id as get_paper lists it: fig1, fig2, extfig1.
panel requiredA panel id as get_paper lists it: a, b, d-inset.
filedata.csv (default), or another file get_panel lists, such as overlay.csv or meta.json.
columnsOnly these columns, by name, in this order. For wide files.
offsetRows to skip, default 0. Pass the previous answer's next_offset to continue.
max_rowsRows per answer, default 500, at most 5,000.

Ask: "Plot Fig. 1(e) of LCP000034 from all of its data." The assistant might call it with:

{"paper_id":"LCP000034","figure":"fig1","panel":"e"}

cite

A citation for the paper. Its note says where the data is, with the package version: the paper's LibCorpus page, or one panel's page when a figure and panel are given.

cite(paper_id, figure?, panel?, format?)
paper_id requiredA LibCorpus id, such as LCP000034. search_papers turns a DOI or an arXiv id into one.
figureWith panel: name that panel's page in the note.
panelWith figure: the panel id.
formattext (default), bibtex or csl.

Ask: "Give me a BibTeX entry for Fig. 1(e) of LCP000034." The assistant might call it with:

{"paper_id":"LCP000034","figure":"fig1","panel":"e","format":"bibtex"}

For developers

The server speaks MCP over Streamable HTTP: JSON-RPC 2.0 posted to the address, with no session and no sign-in, so a request needs no initialize first. The Accept header must name both JSON and event streams; answers come back as plain JSON. List the tools, then call one:

curl -X POST https://libcorpus.org/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
curl -X POST https://libcorpus.org/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":2,"method":"tools/call",
      "params":{"name":"search_papers","arguments":{"query":"Rydberg"}}}'

These use the quoting of macOS, Linux and Git Bash terminals. In Windows PowerShell, call curl.exe and put the body in a file: -d '@body.json'. The endpoint's status codes are under MCP server below.

papers

The corpus and one paper at a time.

GET/api/v1

Entry point

200Links to the corpus, the search and the documentation. application/json
curl https://libcorpus.org/api/v1

GET/api/v1/papers

List the papers in the corpus Newest first, by the date each paper joined LibCorpus. since keeps the papers added on or after a date, so a mirror can ask for what is new since it last looked.

since queryOnly papers added on or after this date (YYYY-MM-DD).
limit queryItems per page, default 100.
offset queryItems to skip; next gives the following page ready-made.
200A page of papers. application/json PaperList
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
curl https://libcorpus.org/api/v1/papers

GET/api/v1/papers/{paperId}

One paper with every figure and panel The paper's summary (as in the list) plus its abstract, journal, citation, the hosted PDF, and each figure with its panels and their descriptions. Use this first; then fetch the panel record for its columns and files.

paperId path, requiredLibCorpus paper id, for example LCP000001.
200The paper. application/json Paper
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
404No such paper, panel, file or address. application/json Error
curl https://libcorpus.org/api/v1/papers/LCP000001

GET/api/v1/papers/{paperId}/paper.pdf

The paper's PDF The PDF LibCorpus hosts, the version named by the paper's primary_source (the paper record's pdf gives its version, size and sha256). Only hosted copies are served.

paperId path, requiredLibCorpus paper id, for example LCP000001.
200PDF. application/pdf
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
404No such paper, panel, file or address. application/json Error
curl https://libcorpus.org/api/v1/papers/LCP000001/paper.pdf

GET/api/v1/papers/{paperId}/data.zip

Every panel's data in one zip Each hosted panel's data files as the file addresses serve them by default (a CSV with its provenance header), at <paperId>/<fig>/<panel>/<name>, and the paper's citation.bib at <paperId>/. The plotting metadata is not included; read it per panel. Figures shown only in the PDF have no files.

paperId path, requiredLibCorpus paper id, for example LCP000001.
200Zip archive. application/zip
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
404No such paper, panel, file or address. application/json Error
curl https://libcorpus.org/api/v1/papers/LCP000001/data.zip

GET/doi/{doi}

Resolve a DOI to its paper page Redirects (302) to /paper/{paperId}. /arxiv/{arxivId} does the same for an arXiv identifier, with or without a version suffix.

doi path, requiredThe bare DOI, for example 10.1038/s41534-026-01329-5.
302Redirect to the paper page.
404No paper with this DOI (the site's not-found page).
curl https://libcorpus.org/doi/10.1038/s41534-026-01329-5

GET/api/health

What this instance serves

200Status, data branch and commit, paper count. application/json
curl https://libcorpus.org/api/health

panels

One panel: what it plots, its columns, its files.

GET/api/v1/papers/{paperId}/figures/{fig}/panels/{panel}

One panel: columns, series, files, citation What is needed to plot the panel: its description and plot type, the columns with their roles (x, y, z, error, band) and the symbols as plotted, for series panels which columns form each curve, and every file with its layout, size and hashes. The panel's meta.json (the site's plotting recipe: colours, ranges, line styles) is one of the files.

paperId path, requiredLibCorpus paper id, for example LCP000001.
fig path, requiredFigure id as in the paper's figure list: fig1, fig2, extfig1.
panel path, requiredPanel id within the figure: a, b, c, or a compound id such as d-inset.
200The panel. application/json Panel
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
404No such paper, panel, file or address. application/json Error
curl https://libcorpus.org/api/v1/papers/LCP000001/figures/fig2/panels/a

files

The panel's files as files: CSV, JSON, SVG.

GET/api/v1/papers/{paperId}/figures/{fig}/panels/{panel}/files/{file}

One file of the panel data.csv holds the plotted values; overlay.csv, when there is one, a fit or model drawn over them; meta.json the plotting recipe. A CSV's layout (wide, long or series) is given in the panel record, and it begins with # comment lines naming the panel and its page URL, the paper, the version LibCorpus hosts with its licence (and the published version when that differs), the data's licence or terms, the package version and the provenance route; readers that honour a comment character skip them (pandas read_csv(url, comment="#"), R read.csv(url, comment.char="#"), numpy loadtxt). JSON comes without that header unless asked for.

paperId path, requiredLibCorpus paper id, for example LCP000001.
fig path, requiredFigure id as in the paper's figure list: fig1, fig2, extfig1.
panel path, requiredPanel id within the figure: a, b, c, or a compound id such as d-inset.
file path, requiredFile name as listed in the panel record's files.
header query1 adds the provenance header, 0 leaves it out. Default: 1 for CSV, 0 for JSON. With 0 a CSV is the stored file byte for byte, which is what data_sha256 is of.
200The file, with its own content type. text/csv, application/json, image/svg+xml
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
404No such paper, panel, file or address. application/json Error
curl https://libcorpus.org/api/v1/papers/LCP000001/figures/fig2/panels/a/files/data.csv

Find a paper or a panel.

Search papers and panel descriptions Ranks like the site's search box: an identifier (LCP id, DOI, arXiv id, with or without prefix or link) first, then title, authors, abstract, then panel descriptions; newest first within a rank. Each result lists up to five of its matching panels and says how many matched.

q query, requiredWords, or an identifier.
limit queryItems per page, default 20.
offset queryItems to skip; next gives the following page ready-made.
200A page of ranked results. application/json SearchPage
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
curl https://libcorpus.org/api/v1/search?q=Rydberg

citations

Ready-made citations.

GET/api/v1/papers/{paperId}/citation.bib

BibTeX entry for the paper An @article (journal) or @misc (arXiv) entry whose note names the paper's LibCorpus page and package version: Figure data (package v3): \url{...}.

paperId path, requiredLibCorpus paper id, for example LCP000001.
200One BibTeX entry. application/x-bibtex
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
404No such paper, panel, file or address. application/json Error
curl https://libcorpus.org/api/v1/papers/LCP000001/citation.bib

GET/api/v1/papers/{paperId}/citation.json

CSL-JSON record for the paper For Zotero, Mendeley, Pandoc and citeproc; note names the LibCorpus page and package version.

paperId path, requiredLibCorpus paper id, for example LCP000001.
200A CSL-JSON item. application/vnd.citationstyles.csl+json
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
404No such paper, panel, file or address. application/json Error
curl https://libcorpus.org/api/v1/papers/LCP000001/citation.json

GET/api/v1/papers/{paperId}/figures/{fig}/panels/{panel}/citation.bib

BibTeX entry naming one panel The paper's entry, with the note naming this panel's page instead: Figure data (package v3): Fig. 2(a), \url{...}. /citation.json beside it gives the same as CSL-JSON.

paperId path, requiredLibCorpus paper id, for example LCP000001.
fig path, requiredFigure id as in the paper's figure list: fig1, fig2, extfig1.
panel path, requiredPanel id within the figure: a, b, c, or a compound id such as d-inset.
200One BibTeX entry. application/x-bibtex
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
404No such paper, panel, file or address. application/json Error
curl https://libcorpus.org/api/v1/papers/LCP000001/figures/fig2/panels/a/citation.bib

GET/api/v1/papers/{paperId}/figures/{fig}/panels/{panel}/citation.json

CSL-JSON record naming one panel

paperId path, requiredLibCorpus paper id, for example LCP000001.
fig path, requiredFigure id as in the paper's figure list: fig1, fig2, extfig1.
panel path, requiredPanel id within the figure: a, b, c, or a compound id such as d-inset.
200A CSL-JSON item. application/vnd.citationstyles.csl+json
400A malformed id (invalid_id) or a missing, unknown or out-of-range parameter (invalid_parameter). application/json Error
404No such paper, panel, file or address. application/json Error
curl https://libcorpus.org/api/v1/papers/LCP000001/figures/fig2/panels/a/citation.json

text

Plain-text and Markdown views for language models.

GET/llms.txt

Guide for language models What LibCorpus is, how to read a paper and fetch a panel, and the list of papers with links to their Markdown digests. /llms-full.txt holds every paper's digest in one file.

200Plain text. text/plain
curl https://libcorpus.org/llms.txt

GET/paper/{paperId}.md

One paper as Markdown Metadata, abstract and every figure and panel with its description, columns and file URLs. The HTML address /paper/{paperId} answers the same when asked with Accept: text/markdown. A panel has its own digest at /paper/{paperId}/fig/{fig}/panel/{panel}.md.

paperId path, requiredLibCorpus paper id, for example LCP000001.
200Markdown. text/markdown
404Not found.
curl https://libcorpus.org/paper/LCP000001.md

mcp-server

The endpoint assistants call. How to connect, and what each tool does: Connect an AI assistant and MCP tools, above.

POST/mcp

MCP server (Model Context Protocol, Streamable HTTP) LibCorpus as a tool for AI assistants. Stateless JSON-RPC over POST, no authentication; GET and DELETE answer 405, except that a browser opening the address is redirected to /api/docs#connect. Tools: search_papers, list_papers, get_paper (JSON or Markdown; one figure with figure), get_panel, get_panel_data (CSV text, whole rows up to max_rows from offset; chosen columns with columns), cite (text, BibTeX, CSL-JSON; naming one panel with figure and panel). Every answer is held under about 40,000 characters so it fits an assistant's context; a shortened answer says what it left out and gives the address of the whole file or record. Search results, the paper list and a panel's rows come in pages, with total, next_offset and complete in the tools' structured results; search_papers, list_papers and get_panel_data declare output schemas. Add the address as a remote MCP server in Claude, ChatGPT, Cursor or any MCP client. Each tool, its arguments and an example: https://libcorpus.org/api/docs#tools.

200The JSON-RPC response. application/json
400A body that is not JSON: a JSON-RPC parse error (-32700), id null. application/json
403A browser Origin other than this site or localhost: a JSON-RPC error, id null. Clients that send no Origin (assistants, editors, scripts) are not affected; pages in other origins read the same data through /api/v1. application/json
curl -X POST https://libcorpus.org/mcp \
  -H 'Content-Type: application/json' -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Schemas

Error

fieldtype
errorobject

LicenseRef

fieldtype
namestring | null e.g. CC BY 4.0
urlstring | null
termsstring | nullSet on data only, for data the authors provided without a licence of its own: the terms it is hosted under, in place of a licence (name and url are then null). Null when the data has a licence.

License

fieldtype
paperLicenseRef
dataLicenseRef
data_differs_from_paperbooleanTrue when the data carries terms different from the article's licence: a deposit licence of its own, or terms in place of a licence.

PaperSummary

fieldtype
idstring e.g. LCP000001
titlestring | null
authorsarray of string
yearinteger | nullPublication year.
venuestring | nullJournal abbreviation, or arXiv when the hosted version is the preprint.
primary_sourcestring | nullThe version LibCorpus hosts and the data follows. For a published paper whose journal version cannot be hosted this is arxiv, with doi still set.
doistring | nullAt least one of doi and arxiv_id is set.
arxiv_idstring | nullBare arXiv id, without version.
arxiv_versionstring | nullThe pinned arXiv version, v1, v2...; null exactly when arxiv_id is. With primary_source arxiv, the version the data was checked against.
licenseLicense
data_package_versioninteger | nullBumped when the package is re-released; printed in every CSV header and citation note.
figure_countinteger | nullFigures with at least one data-backed panel.
panel_countinteger | nullData-backed panels.
data_pointsinteger | nullValues across the paper's CSVs.
date_addedstring | nullWhen the paper joined LibCorpus; the list's order and since use it.
linksobjectAbsolute URLs of related resources.

PaperList

fieldtype
totalintegerHow many match in all.
countintegerHow many are in this page.
offsetintegerPosition of this page's first item in the whole list.
nextstring | nullThe following page, or null on the last.
itemsarray of PaperSummary

Paper

fieldtype
idstring e.g. LCP000001
titlestring | null
authorsarray of string
yearinteger | nullPublication year.
venuestring | nullJournal abbreviation, or arXiv when the hosted version is the preprint.
primary_sourcestring | nullThe version LibCorpus hosts and the data follows. For a published paper whose journal version cannot be hosted this is arxiv, with doi still set.
doistring | nullAt least one of doi and arxiv_id is set.
arxiv_idstring | nullBare arXiv id, without version.
arxiv_versionstring | nullThe pinned arXiv version, v1, v2...; null exactly when arxiv_id is. With primary_source arxiv, the version the data was checked against.
licenseLicense
data_package_versioninteger | nullBumped when the package is re-released; printed in every CSV header and citation note.
figure_countinteger | nullFigures with at least one data-backed panel.
panel_countinteger | nullData-backed panels.
data_pointsinteger | nullValues across the paper's CSVs.
date_addedstring | nullWhen the paper joined LibCorpus; the list's order and since use it.
linksobjectAbsolute URLs of related resources.
abstractstring | null
journalJournal
citationobject
pdfPdf
figuresarray of Figure

Journal

fieldtype
namestring e.g. npj Quantum Information
abbreviationstring | null e.g. npj Quantum Inf.
volumestring | null
pagesstring | null

Pdf

fieldtype
urlstring
versionstring | nullWhich version it is. e.g. arXiv:2309.06324v1, npj Quantum Inf. (2026), DOI 10.1038/s41534-026-01329-5
bytesinteger
sha256string

Figure

fieldtype
idstring e.g. fig2
labelstring e.g. Fig. 2
statusstringdata when panels are present; otherwise no_data, data_unavailable, data_ambiguous or pending (the figure is in the PDF only).
pdf_pageinteger | nullThe page of the hosted PDF the figure is printed on, 1-based; null where not recorded.
panelsarray of PanelSummary

PanelSummary

fieldtype
idstring e.g. a
typestring | nullplot or schematic.
plot_typestring | null e.g. series_1d, heatmap_2d, layered_2d
representationstringdata (CSV-backed), image (SVG with a description) or structured.
descriptionstring | nullWhat the panel plots.
linksobjectAbsolute URLs of related resources.

FileRef

fieldtype
namestring e.g. data.csv, overlay.csv, meta.json
kindstring e.g. data, overlay, meta, image
labelstring | nullWhat a secondary file holds, in the paper's words where its meta layer is named. e.g. analytical sol.
layoutstring | nullCSV only. wide: one row per x value, one column per curve sharing it. long: one row per grid point (heatmaps, surfaces). series: each curve names its own x and y columns, listed in the panel's series.
urlstring
bytesinteger | nullSize of the file as a default download returns it (CSV with its header, JSON without).
sha256string | nullHash of the file as a default download returns it.
data_sha256string | nullCSV only: hash of the values alone, the stored file, which ?header=0 returns. Changes only when the values do, so it is the one to cite.

Column

fieldtype
namestringCSV column name.
rolestring | null e.g. x, y, z, error, band lower, band upper, region x, region y
symbolstring | nullThe quantity as plotted, LaTeX where the paper uses it.
filestringWhich file the column is in.

Series

fieldtype
namestring | nullThe curve's legend label, LaTeX where the paper uses it.
partstring | nullregion: x and y are the vertices of a filled closed outline (a shaded area), NaN rows separating polygons.
filestring
xstring | null
ystring | null
errorstring | null
error_plusstring | null
error_minusstring | null
band_lowerstring | null
band_upperstring | null

Panel

fieldtype
paperPaperSummary
figureobject
panelobject
columnsarray of Column
seriesarray | nullSeries panels only: which columns form each curve. Null otherwise.
filesarray of FileRef
citationobject
linksobjectAbsolute URLs of related resources.

SearchResult

fieldtype
idstring e.g. LCP000001
titlestring | null
authorsarray of string
yearinteger | nullPublication year.
venuestring | nullJournal abbreviation, or arXiv when the hosted version is the preprint.
primary_sourcestring | nullThe version LibCorpus hosts and the data follows. For a published paper whose journal version cannot be hosted this is arxiv, with doi still set.
doistring | nullAt least one of doi and arxiv_id is set.
arxiv_idstring | nullBare arXiv id, without version.
arxiv_versionstring | nullThe pinned arXiv version, v1, v2...; null exactly when arxiv_id is. With primary_source arxiv, the version the data was checked against.
licenseLicense
data_package_versioninteger | nullBumped when the package is re-released; printed in every CSV header and citation note.
figure_countinteger | nullFigures with at least one data-backed panel.
panel_countinteger | nullData-backed panels.
data_pointsinteger | nullValues across the paper's CSVs.
date_addedstring | nullWhen the paper joined LibCorpus; the list's order and since use it.
linksobjectAbsolute URLs of related resources.
matchstring
matching_panel_countintegerHow many of the paper's panel descriptions match.
matching_panelsarray of objectUp to five of them.

SearchPage

fieldtype
querystring
totalintegerHow many match in all.
countintegerHow many are in this page.
offsetintegerPosition of this page's first item in the whole list.
nextstring | nullThe following page, or null on the last.
itemsarray of SearchResult