biorxiv-mcp-server
io.github.cyanheads/biorxiv-mcp-servercyanheadsstreamable-httpv0.6.0Scanned
Search and retrieve bioRxiv and medRxiv preprints — by DOI, date interval, or keyword — via MCP.
7 tools: 7 only read.
- Reads private data · 0
- Reads text others write · 6
- Can send data out · 0
- 2 findings
Tools listed without sign-in. Remote posture
With Anlyon
Decide which of this server's tools your agent can use.
- Anlyon lists the tools and adopts none.
- Each tool you adopt gets approvals and receipts.
- Your agent never holds the key.
A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.
What the biorxiv-mcp-server can do
The biorxiv-mcp-server is published by the GitHub account cyanheads. It has 7 tools. Every one of them only reads: going by their definitions, none of them creates, changes, sends or deletes anything.
6 tools bring in text other people can write (biorxiv_list_recent, biorxiv_get_preprint, biorxiv_get_published_version and 3 more).
The fixed checks matched 2 findings (tool-shadowing). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in, and its tools changed 3 times since Anlyon Radar first saw them.
Last scanned Nov 21, 2026, version 0.6.0.
Tools
Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.
7 tools
Only reads 7Show 7 read-only tools
biorxiv_get_fulltextRetrieve a preprint's full text as best-effort Markdown, extracted from its rendered HTML article page.Reads text others writeNo findings
Why it is in this group
- The name "biorxiv_get_fulltext" says "get": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe description's first sentence says "retrieve" and "html": it reads web pages, which anyone can write.
Description, as the server wrote it Retrieve a preprint's full text as best-effort Markdown, extracted from its rendered HTML article page. Reads the latest version unless one is requested (the version input, or a vN suffix on the DOI), confirms it via the details API, then fetches and extracts the body — abstract, sections, inline citation labels, and the reference list. Table cells are not on the article page: each table appears as its caption, and omittedTables names it with the URL that serves its cells. bioRxiv and medRxiv share the 10.1101/ DOI prefix, so server="both" (the default) resolves the DOI against both in parallel and the response reports which server answered. This is HTML-to-Markdown extraction, not structured JATS: section structure is approximate and not guaranteed. Long articles exceed a single response, so use offset and limit to page through them (the response reports totalChars, remainingChars, and hasMore); paging is cheap because the extracted article is cached per version for an hour after the first read, so only the first chunk pays for a fetch. Not every preprint has an extractable HTML page — some are PDF-only and some origins block programmatic access — in which case a fulltext_unavailable error routes you to biorxiv_get_preprint for the title, metadata, and current abstract. For a preprint that has been published in a journal, the journal's version may have richer full text elsewhere.
Annotations the server declares
- readOnlyHint: true
- openWorldHint: true
biorxiv_get_preprintFetch full metadata, the abstract, all revision history, JATS XML full-text links, and published-journal DOI for one or more preprints by DOI.Reads text others writeNo findings
Why it is in this group
- The name "biorxiv_get_preprint" says "get": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Fetch full metadata, the abstract, all revision history, JATS XML full-text links, and published-journal DOI for one or more preprints by DOI. Each DOI returns all revisions in one response, with the abstract on the latest revision only — the API keeps one abstract per preprint. Each call looks up the first 10 distinct DOIs and returns the rest, never looked up, in unprocessed[] — pass that list as dois in another call to continue; a repeated DOI is looked up once and an unparseable entry is reported in failed[], so neither counts toward the 10. When server="both" (default), each DOI is checked against both bioRxiv and medRxiv; the response includes which server the preprint was found on. Failed lookups are reported per-DOI in failed[] rather than aborting the batch, each carrying a reason (not_found, invalid_doi_format, upstream_unavailable, rate_limited) and a retryable flag; a rate_limited entry also carries the wait in seconds the origin asked for. preprints[], failed[], and unprocessed[] follow input order. DOIs must match the pattern 10.NNNN/…; a doi.org or article URL, a doi: label, and a trailing vN / .full suffix are stripped first, and results report the bare DOI.
Annotations the server declares
- readOnlyHint: true
- idempotentHint: true
- openWorldHint: true
biorxiv_get_published_versionResolve a DOI to its preprint-to-journal crosswalk record — preprint DOI, journal DOI, journal name, published date, and corresponding author details — from either side: a preprint…Reads text others writeNo findings
Why it is in this group
- The name "biorxiv_get_published_version" says "get": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Resolve a DOI to its preprint-to-journal crosswalk record — preprint DOI, journal DOI, journal name, published date, and corresponding author details — from either side: a preprint DOI (10.1101/ or 10.64898/) finds the journal article it was published as, and any other DOI is read as a journal article's and finds the bioRxiv or medRxiv preprint it was posted as. Use when biorxiv_get_preprint shows a `publishedJournalDoi` and you need the journal details, or when you hold a journal article and want its preprint. bioRxiv and medRxiv share their DOI prefixes, so server="both" (the default) checks both in parallel and the response reports which server answered. When the crosswalk holds no record for a published preprint, the journal DOI still comes back from the preprint's own record, without journal name or date, and a notice says so. Returns a not-found error when no server holds the preprint, it lists no journal version, or no preprint is recorded for the journal DOI.
Annotations the server declares
- readOnlyHint: true
- idempotentHint: true
- openWorldHint: true
biorxiv_list_categoriesList valid subject category strings for bioRxiv and medRxiv — the categories the listing API actually filters on.No findings
Why it is in this group
- The name "biorxiv_list_categories" says "list": a word for reading.
Description, as the server wrote it List valid subject category strings for bioRxiv and medRxiv — the categories the listing API actually filters on. Use these strings as the `category` filter in biorxiv_list_recent to narrow results to a specific field; case, and "_" or "-" in place of a space, do not matter there. Run this tool before filtering to get the current valid values.
Annotations the server declares
- readOnlyHint: true
- idempotentHint: true
- openWorldHint: false
biorxiv_list_published_versionsList preprints that appeared as journal articles within a journal publication-date interval, each with its journal DOI, journal name, publication date, and preprint metadata — the …Reads text others write1 finding
Why it is in this group
- The name "biorxiv_list_published_versions" says "list": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Warn
The description contains "other server", wording about other tools or servers.
What this check looks forMatched text
description: "...ver="both" does not abort the call: the other server's page is still returned and the failed..."
Description, as the server wrote it List preprints that appeared as journal articles within a journal publication-date interval, each with its journal DOI, journal name, publication date, and preprint metadata — the same record biorxiv_get_published_version returns for one DOI. The dates filter on journal publication, not preprint posting, so most rows were posted months or years earlier; use biorxiv_list_recent for preprints posted in an interval. `publisher`, a journal DOI prefix such as 10.7554 for eLife, limits the listing to the bioRxiv preprints published under it; those rows carry no journal name, authors, or abstract, and a prefix api.biorxiv.org does not recognize raises invalid_publisher. Returns `limit` publications per server (default 25, up to 100 — the page api.biorxiv.org serves; a larger limit returns 100 and says so). Continue at the `nextCursor` a page returns rather than stepping `cursor` by a fixed amount; `limit: 100` walks an interval in the fewest calls. When server="both" (default), per-server pagination state is returned separately, since one cursor can be past the end of one server and inside the other — to continue a single server, pass its `nextCursor` with `server` set to that server. One server failing under server="both" does not abort the call: the other server's page is still returned and the failed one is named in `failed[]`. Every attempted server failing raises a retryable upstream_unavailable (or rate_limited) error rather than an empty page. Abstracts are omitted by default to keep the page small — pass include_abstract: true for the whole page. Every preprintDoi, 10.1101/ and 10.64898/ alike, is accepted by biorxiv_get_preprint and the other DOI tools.
Annotations the server declares
- readOnlyHint: true
- openWorldHint: true
biorxiv_list_recentList preprints posted or revised within a date interval, optionally scoped to one server or a subject category.Reads text others write1 finding
Why it is in this group
- The name "biorxiv_list_recent" says "list": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Warn
The description contains "other server", wording about other tools or servers.
What this check looks forMatched text
description: "...ver="both" does not abort the call: the other server's page is still returned and the failed..."
Description, as the server wrote it List preprints posted or revised within a date interval, optionally scoped to one server or a subject category. Returns 30 preprints per page (fixed by the API); pass `cursor` as an integer offset (0, 30, 60, …) to step through additional pages. Abstracts are omitted by default to keep the page small — pass include_abstract: true for the whole page (each row carries the latest-revision abstract of its preprint, whichever revision the row is), or call biorxiv_get_preprint (up to 10 DOIs per call) for a few. When server="both" (default), per-server pagination state is returned separately — use each server's `cursor` field for independent advancement. One server failing under server="both" does not abort the call: the other server's page is still returned and the failed one is named in `failed[]`, marking the result set as partial rather than complete. Every attempted server failing is a different case and does abort the call, with a retryable upstream_unavailable (or rate_limited) error — an empty page would otherwise be indistinguishable from an interval that genuinely holds nothing. Call biorxiv_list_categories for valid category strings; a server that answers a category filter with its unfiltered listing is left out with a notice, and invalid_category is raised when no server applied it. `funder` limits the listing to bioRxiv preprints funded by that organization, given its ROR ID; an ID api.biorxiv.org has no funder record for raises invalid_funder rather than returning an empty page.
Annotations the server declares
- readOnlyHint: true
- openWorldHint: true
biorxiv_search_preprintsSearch preprints by keyword and/or author using EuropePMC for relevance ranking, then enrich matching DOIs with full bioRxiv/medRxiv metadata.Reads text others writeNo findings
Why it is in this group
- The name "biorxiv_search_preprints" says "search": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Search preprints by keyword and/or author using EuropePMC for relevance ranking, then enrich matching DOIs with full bioRxiv/medRxiv metadata. Provide a keyword query, an author name, or both — author maps to an EuropePMC AUTH: field query and is ANDed with the keyword query. Covers both servers by default. EuropePMC indexes new preprints within 1–2 days of posting; for preprints posted within the last day, prefer biorxiv_list_recent. The default page requests 10 results with abstracts. Set include_abstract: false to omit them, or use biorxiv_get_preprint for abstracts for up to 10 DOIs per call. A EuropePMC rate limit (HTTP 429) fails the call with a retryable rate_limited error carrying the wait in seconds — a rate-limited metadata enrichment does not, and instead marks the affected record enrichment_error: "rate_limited".
Annotations the server declares
- readOnlyHint: true
- openWorldHint: true
This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.
Exfil surface
Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.
No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data or outbound.
- Reads private data0
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write6
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out0
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.
Remote posture
What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.
Looked at Nov 20, 2026, 19:24 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.
- Reachable
- yes
- HTTP status of an anonymous initialize
- 200
- tools/list served without credentials
- yes
- Tools that change things, listed without credentials
- 0
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- not observed
- Authorization servers
- not observed
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- not observed
- Client ID metadata documents
- not observed
Findings
Matches of fixed checks, with the text they matched.
The description contains "other server", wording about other tools or servers.
What this check looks forMatched text
description: "...ver="both" does not abort the call: the other server's page is still returned and the failed..."
The description contains "other server", wording about other tools or servers.
What this check looks forMatched text
description: "...ver="both" does not abort the call: the other server's page is still returned and the failed..."
History
Changes to its tool definitions and to its remote endpoint, newest first.
Tools1 changed.~1 changedView diff
- Changed:biorxiv_search_preprintsOnly readsdescriptioninputSchema
Tools1 changed.~1 changedView diff
- Changed:biorxiv_search_preprintsOnly readsinputSchema
Tools5 changed.~5 changedView diff
- Changed:biorxiv_get_fulltextOnly readsdescription
- Changed:biorxiv_get_preprintOnly readsdescription
- Changed:biorxiv_get_published_versionOnly readsdescriptioninputSchema
- Changed:biorxiv_list_published_versionsOnly readsdescriptioninputSchema
- Changed:biorxiv_list_recentOnly readsdescriptioninputSchema
Install
Pinned to version 0.6.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http biorxiv https://biorxiv.caseyjhand.com/mcpclaude mcp add --transport http biorxiv https://biorxiv.caseyjhand.com/mcpRun in a terminal.
Claude Code permissions
Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.
{
"permissions": {
"allow": [
"mcp__biorxiv__biorxiv_list_categories",
"mcp__biorxiv__biorxiv_list_recent",
"mcp__biorxiv__biorxiv_get_preprint",
"mcp__biorxiv__biorxiv_get_published_version",
"mcp__biorxiv__biorxiv_list_published_versions",
"mcp__biorxiv__biorxiv_search_preprints",
"mcp__biorxiv__biorxiv_get_fulltext"
],
"ask": [],
"deny": []
}
}Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as biorxiv. Review it before you use it.
Capability tags behind 6 rules
- mcp__biorxiv__biorxiv_list_recent untrusted-input
- mcp__biorxiv__biorxiv_get_preprint untrusted-input
- mcp__biorxiv__biorxiv_get_published_version untrusted-input
- mcp__biorxiv__biorxiv_list_published_versions untrusted-input
- mcp__biorxiv__biorxiv_search_preprints untrusted-input
- mcp__biorxiv__biorxiv_get_fulltext untrusted-input
Pin
Keep the decision: fail CI when the tools change, follow every change, or show a badge.
Pin in CI
Fail a build when the tools you reviewed are not the tools the server serves today.
curl -s "https://radar.anlyon.com/api/v1/servers/io.github.cyanheads/biorxiv-mcp-server/verify?hash=291a2e1ed77b677200eb7f04d980ebe40c107adf92f180d70edc575bcd8f6f19" | jq -e '.changed == false'Pinned hash 291a2e1ed77b, from Nov 21, 2026, 18:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.cyanheads/biorxiv-mcp-server/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.cyanheads/biorxiv-mcp-server)Source: MCP Registry · repository. Synced Nov 21, 2026, 18:17 UTC.
biorxiv-mcp-server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the biorxiv-mcp-server safe?
- No scan can promise that. What its definitions show: 7 tools, all of which only read, and 2 findings from fixed checks. Read-only tools can still pass private data to a model.
Can the biorxiv-mcp-server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the biorxiv-mcp-server do things that cannot be undone?
- No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.
What tools does the biorxiv-mcp-server have?
- 7 tools: 7 read. biorxiv_list_categories, biorxiv_list_recent, biorxiv_get_preprint, biorxiv_get_published_version, biorxiv_list_published_versions, biorxiv_search_preprints and biorxiv_get_fulltext.
Does the biorxiv-mcp-server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the biorxiv-mcp-server send data to outside services?
- None of its tools sends, posts or publishes, or takes a recipient, channel or URL.
Has the biorxiv-mcp-server changed its tools?
- Yes, 3 times since Anlyon Radar started tracking it, most recently on Nov 21, 2026. Every change is on this page and in its Atom feed.