Provides tools for searching academic papers, retrieving paper metadata, citations, references, related papers, and recommendations using the Semantic Scholar API.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@sophotron search for papers on the nature of consciousness"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
sopho- (σοφία, wisdom) + -tron (instrument) — a wisdom-instrument for philosophy research.
Sophotron is a tool for searching and retrieving philosophy papers, built for AI research assistants that need to ground themselves in the actual philosophical literature rather than the shallow, mis-classified philosophy slice of general academic databases. It is exposed today as an MCP server; a CLI and other frontends can sit on the same interface-agnostic core.
It combines four open sources, each doing what it does best:
Source | Role |
Semantic Scholar | Search, metadata, citations, references, SPECTER-based related papers (the backbone) |
OpenAlex | Open-access discovery — agent-fetchable PDFs, ~11.5M OA philosophy works |
Unpaywall | DOI → best open-access PDF location |
CrossRef | Authoritative DOI metadata fallback ( |
PhilArchive | Philosophy-specific open-access cache (metadata + best-effort PDF links) |
PhilPapers taxonomy | Browse 6,134 expert-curated philosophy categories (cached snapshot) |
Runs fully keyless. An optional Semantic Scholar API key raises rate limits; optional contact emails opt into the OpenAlex/Unpaywall "polite pool." No source requires registration.
Why four sources? Philosophy is under-served by general infrastructure: low open-access (~15%) and DOI-assignment (~60%) rates, canonical works predating DOIs, and systematic mis-classification (Rawls's A Theory of Justice tagged "Sociology" in Semantic Scholar; Heidegger's Being and Time with a null field). And as of 2026, the PhilPapers/PhilArchive web APIs sit behind Cloudflare bot-protection. OpenAlex + Unpaywall therefore supply the open-access PDFs an automated client can actually fetch, while a cached PhilArchive harvest adds philosophy-specific coverage.
npm install
npm run build
npm run smoke # optional: MCP stdio smoke testRequires Node.js ≥ 22.
Related MCP server: CogniResearch
All configuration is optional — copy .env.example to .env and fill in what you want:
# Optional: raises Semantic Scholar rate limits (request at
# https://www.semanticscholar.org/product/api). Without it, S2 still works but
# rate-limits aggressively (HTTP 429) on its shared unauthenticated pool.
SEMANTIC_SCHOLAR_API_KEY=
# Optional: OpenAlex/Unpaywall "polite pool" contact (a courtesy id, not a key).
# OPENALEX_EMAIL=you@example.com
# UNPAYWALL_EMAIL=you@example.comAdvanced overrides (SOPHOTRON_OAI_ENDPOINT, SOPHOTRON_DATA_DIR, SOPHOTRON_LIBRARY_PATH) are documented in .env.example.
Claude Code / any stdio MCP client (~/.claude/settings.json):
{
"mcpServers": {
"sophotron": {
"command": "node",
"args": ["/absolute/path/to/sophotron/dist/index.js"],
"env": { "SEMANTIC_SCHOLAR_API_KEY": "optional" }
}
}
}Claude Desktop / Cowork (MCPB bundle): a manifest.json is included. Pack it with the MCPB CLI and install the resulting .mcpb:
npx @anthropic-ai/mcpb packTool | Description |
| Search via Semantic Scholar (auto-falls back to OpenAlex when S2 is rate-limited); PhilArchive PDF enrichment. Optional |
| Search via OpenAlex with open-access PDF discovery ( |
| Paper details by S2 ID, DOI, arXiv ID, or PhilArchive code |
| Papers citing a paper |
| A paper's bibliography |
| SPECTER-based related papers. Optional |
| Resolve an open-access PDF via the fallback chain (below) |
| Search the local PhilArchive open-access cache |
| Incrementally harvest the PhilPapers/PhilArchive OAI-PMH feed into the cache |
| PhilArchive cache + cross-reference statistics |
| Browse the expert-curated PhilPapers category taxonomy (6,134 categories) |
| Papers near an inferred taxonomy-category region, ranked by influence — a semantic lens over an inferred mapping, not an authoritative index |
| Resolve an open-access PDF, extract its text, and store it as markdown (arxiv-mcp-server-compatible path) |
| Read back the stored text of a downloaded paper, paginated |
| A personal paper library (JSON-on-disk) with tags and notes |
| "More like these," seeded by your library, via Semantic Scholar's hosted Recommendations API (SPECTER2) |
The PhilPapers taxonomy is more than a list to browse: each category is treated as an inferred
SPECTER2 region (a coordinate in embedding space, seeded from the category name), so you can
navigate papers relative to a category. Pass a category_id (the id browse_taxonomy returns)
to search_papers or get_related to rerank the results by proximity to that region, or call
category_papers for influence-ranked papers near it.
This is a semantic lens over an inferred mapping, not a classification claim — results are
tagged with relative proximity (nearer / farther / unplaceable), placement: "inferred", and an
honest confidence tier (distinct for semantically-clear coarse categories, low for fine,
catch-all, or thinly-seeded ones). Transient Semantic Scholar outages are reported as
unavailable, never as a confident empty region.
get_pdf)Tries sources in order of reliability and stops at the first hit, reporting where the PDF came from and whether an automated client can fetch it:
Semantic Scholar openAccessPdf → agentFetchable: true
Unpaywall by DOI → agentFetchable: true
OpenAlex by DOI (or title search) → agentFetchable: true
PhilArchive best-effort URL → agentFetchable: false (philarchive.org is Cloudflare-walled to automated clients; open these in a browser)
// get_pdf { "paper_id": "10.1007/s11019-014-9560-2" }
{ "pdfUrl": "https://…/paper.pdf", "source": "unpaywall", "agentFetchable": true, "oaStatus": "green", … }PhilArchive's OAI-PMH metadata omits DOIs, so papers are linked by fuzzy title + author matching (title 50% / author 35% / year 15%; accept > 0.85). Match results carry a confidence score so uncertain links can be reviewed.
Source | Limit | Handling |
Semantic Scholar | 429s even at slow pace without a key — the unauthenticated pool is globally shared (empirically 7/8 requests 429'd at 5s spacing) | bounded retry/backoff ( |
OpenAlex / Unpaywall | generous, no key | polite-pool email recommended |
PhilArchive OAI ( | 2 req/sec | built-in 500ms throttle |
| Cloudflare 403 to agents | PDF URLs returned as best-effort, browser-only |
src/
├── index.ts # MCP server frontend (18 tools)
├── types.ts # shared types
├── adapters/ # one module per external API; never throw to callers
│ ├── semantic-scholar.ts # S2 Graph/Recommendations API (+ SPECTER2 embeddings)
│ ├── openalex.ts # OpenAlex works + OA discovery
│ ├── unpaywall.ts # DOI → OA PDF
│ ├── crossref-api.ts # api.crossref.org DOI metadata fallback
│ ├── philarchive.ts # PhilArchive/PhilPapers OAI-PMH + cache
│ └── taxonomy.ts # PhilPapers category hierarchy (cached)
├── services/ # cross-source logic
│ ├── crossref.ts # S2 ↔ PhilArchive fuzzy cross-reference
│ ├── pdf-resolver.ts # S2 → Unpaywall → OpenAlex → PhilArchive chain
│ ├── paper-download.ts # download_paper / read_paper full-text store
│ ├── library.ts # personal library (JSON-on-disk)
│ └── taxonomy-bridge.ts # taxonomy→papers semantic bridge (inferred centroids)
└── utils/
├── fuzzy-match.ts # title/author matching
├── vector.ts # cosine / centroid / proximity ranking
├── rate-limit.ts # serial rate limiter (S2 1 req/s)
└── http.ts # shared retry/backoff fetch
data/ # philarchive-cache.json, crossref-cache.json, category-centroids.json (regenerable)
scripts/smoke-test.mjs # MCP stdio smoke test (also runs in CI)npm run build # tsc (tsconfig.build.json — excludes tests/experiments)
npm run typecheck # tsc --noEmit (full, incl. tests)
npm test # vitest (set SKIP_INTEGRATION=true to skip live tests)
npm run smoke # MCP stdio smoke testStart with CONTRIBUTING.md; the operational guide (architecture, conventions, correctness contracts) is CLAUDE.md. Tests are the verification spine — every adapter has unit tests with mocked fetch, plus a network-gated integration suite.
Browse the expert-curated PhilPapers category hierarchy (6,134 categories from a cached snapshot — the live PhilPapers API is Cloudflare-blocked). Omit category_id for the top-level areas; pass category_id to list its subcategories; pass query to find categories by name. With a category_id you can: call category_papers to list influential papers near that category region, or pass it as the category_id param to search_papers / get_related to re-rank by category proximity. (Those are an inferred semantic lens, not an authoritative classification.)
Inputs: limit, query, category_id.
List influential papers near an INFERRED semantic region for a PhilPapers category. This is a semantic lens over an inferred mapping (category name → SPECTER2 centroid → nearby papers), NOT an authoritative PhilPapers classification or a canonical reading list. Pass a category_id from browse_taxonomy. Coarse/distinct categories (top-level areas, subfields) are MORE reliable but still an unvalidated per-category inference; fine-grained leaves — and miscellaneous/structural buckets and thin regions — are flagged low-confidence. Papers are ranked by influential citation count, each tagged with its relative proximity to the region.
Inputs: category_id (required), limit.
Resolve the best open-access PDF for a paper, fetch it, extract its text, and store it as markdown in the shared library (default ~/.arxiv-mcp-server/papers, override with ARXIV_MCP_STORAGE_PATH). Returns a paper_id (slug) to pass to read_paper. PhilArchive PDFs are browser-only (Cloudflare) and are not fetched — the URL is returned to open manually.
Inputs: paper_id (required), title.
Get statistics about the local PhilArchive cache and cross-reference mappings.
Inputs: no named parameters.
Get papers that cite a given paper.
Inputs: paper_id (required), limit.
Get detailed metadata for a paper by ID. Accepts S2 paper ID, DOI, arXiv ID, or PhilArchive ID.
Inputs: paper_id (required).
Get PDF URL for a paper. Tries Semantic Scholar open access first, then falls back to PhilArchive.
Inputs: paper_id (required).
Get papers that a given paper cites (its bibliography).
Inputs: paper_id (required), limit.
Incrementally harvest papers from PhilArchive OAI-PMH. Rate limited to 2 req/sec. Run periodically to update local cache.
Inputs: from_date, max_records.
Save a paper to your personal library (the arxiv-sanity "library" primitive). Resolves metadata from Semantic Scholar (or CrossRef for a bare DOI) and stores it with optional tags and a note. The saved set is what recommend uses to suggest what to read next. Accepts a Semantic Scholar id, DOI, arXiv id, or PhilArchive code.
Inputs: paper_id (required), note, tags.
List the papers in your personal library, newest first, optionally filtered to a tag. Also reports the set of tags in use.
Inputs: tag.
Remove a saved paper from your personal library by id, S2 id, or DOI.
Inputs: paper_id (required).
Read back the markdown text of a paper previously stored by download_paper. Supports pagination via start/max_chars. Reads {storageDir}/{paper_id}.md (shared with arxiv-mcp-server). If the paper is not stored, returns an error asking to download_paper first.
Inputs: paper_id (required), start, max_chars.
Recommend papers to read next, seeded by your library — the arxiv-sanity "more like these" feature. By default it uses your whole library as positive examples; pass tag to recommend from one reading-list bucket, or paper_ids to recommend ad hoc from specific papers. Powered by Semantic Scholar SPECTER2 embeddings; papers already in your library are excluded. Without a Semantic Scholar API key this may be rate-limited (HTTP 429).
Inputs: tag, limit, paper_ids.
Search papers via OpenAlex with open-access full-text discovery. Complements Semantic Scholar; OpenAlex and the OA PDFs it points to are reachable by automated clients (unlike PhilArchive, which is Cloudflare-walled). Use philosophy_only to restrict to the OpenAlex Philosophy concept, and open_access_only to return only papers with OA full text.
Inputs: query (required), limit, year_to, year_from, +2 more.
Search for philosophy papers using Semantic Scholar. Returns metadata, abstracts, and PDF URLs where available. Pass category_id (from browse_taxonomy) to re-rank results as a semantic lens over an inferred PhilPapers-category region — papers stay under "papers", each tagged with a relative proximity tier ("nearer"/"farther"/"unplaceable"); this is a "near category C" ordering, not an authoritative classification.
Inputs: query (required), limit, offset, year_to, +4 more.
Search the local PhilArchive cache directly. Useful for finding philosophy papers that may not be in Semantic Scholar.
Inputs: query (required), limit.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 18 tools
Each tool has a clearly distinct purpose: multiple search tools are differentiated by backend (Semantic Scholar, PhilArchive, OpenAlex) and use case, while citation, reference, related, PDF retrieval, library management, and taxonomy browsing are all separate actions. No two tools appear to do the same thing.
Tool names mostly follow a verb_noun pattern (get_, search_, browse_, download_, read_), with only a few deviations like 'category_papers' and 'library_add' where the noun comes first, and 'recommend' as a bare verb. Overall the convention is predictable and readable.
With 18 tools, the server is slightly on the heavy side but the breadth of functionality (multi-source search, citation graph, taxonomy, library, full-text reading, recommendations) justifies each tool. No tool feels redundant, though the count approaches the upper bound of what is ideal.
The tool surface covers the full research workflow: discovery (search, browse, related), retrieval (metadata, PDFs, full text), library management (add/remove/list), reading, and recommendation. Minor gaps exist (no bulk operations, no citation export) but none would prevent core use cases.
Academic research MCP server for paper search, citation checks, graphs, and deep research.
Personal knowledge base MCP server with semantic search, auto-categorization, metadata extraction
arXiv MCP — preprint server search (free, no auth)
