OpenArx
This is infrastructure for AI agents doing scientific research at agent speed, not a web app for humans. It exposes 15 specialized search primitives through MCP: fact checking against corpus, methodology lookup, benchmark queries, paper comparison, conceptual landscape mapping. Three profiles scale access: consumer for reading research, publisher for document submission with AI-assisted review, governance for platform decisions. Built because agents are already doing literature reviews and hallucinating citations at scale, and the traditional journal system wasn't designed for that workflow. Open source under Apache 2.0. Reach for this when your agent needs to ground scientific reasoning in actual papers or when you're building research tooling that needs structured access to scientific knowledge instead of scraping search engines.
- tools
- 35
- transport
- HTTP
- last updated
- Jul 29, 2026
Tools
35 toolsPublic tool metadata for what this MCP can expose to an agent.
searchHybrid semantic and keyword search across scientific papers.
Hybrid semantic and keyword search across scientific papers.
No parameter schema in public metadata yet.
search_keywordPure keyword (BM25) search, fastest option for exact-term lookups.
Pure keyword (BM25) search, fastest option for exact-term lookups.
No parameter schema in public metadata yet.
search_semanticPure vector search, best for paraphrased queries and concept exploration.
Pure vector search, best for paraphrased queries and concept exploration.
No parameter schema in public metadata yet.
find_relatedFind related papers by similarity, by entity, or by concept.
Find related papers by similarity, by entity, or by concept.
No parameter schema in public metadata yet.
paginateContinue a previous search from its cached candidate pool instead of re-running it.
Continue a previous search from its cached candidate pool instead of re-running it.
No parameter schema in public metadata yet.
find_evidenceCheck a claim against the corpus and group passages by what supports, contradicts, or is neutral.
Check a claim against the corpus and group passages by what supports, contradicts, or is neutral.
No parameter schema in public metadata yet.
compare_papersSide-by-side comparison of several papers.
Side-by-side comparison of several papers.
No parameter schema in public metadata yet.
explore_topicMap the conceptual landscape around a topic across the paper corpus.
Map the conceptual landscape around a topic across the paper corpus.
No parameter schema in public metadata yet.
find_methodologyFind methodology approaches for a specific research task.
Find methodology approaches for a specific research task.
No parameter schema in public metadata yet.
find_benchmark_resultsQuery structured benchmark scores from papers.
Query structured benchmark scores from papers.
No parameter schema in public metadata yet.
find_codeFind papers with associated code repositories, datasets, or benchmark results.
Find papers with associated code repositories, datasets, or benchmark results.
No parameter schema in public metadata yet.
get_documentRetrieve full paper details by ID.
Retrieve full paper details by ID.
No parameter schema in public metadata yet.
get_chunksRetrieve specific chunks from a known document, filtered by content type, section, or entity.
Retrieve specific chunks from a known document, filtered by content type, section, or entity.
No parameter schema in public metadata yet.
find_by_idResolve an external identifier (OpenArx ID, DOI, arXiv ID, Semantic Scholar Corpus ID, DBLP ID) to a paper.
Resolve an external identifier (OpenArx ID, DOI, arXiv ID, Semantic Scholar Corpus ID, DBLP ID) to a paper.
No parameter schema in public metadata yet.
find_related_claimsFind claims related to an existing claim by that claim's stored vectors.
Find claims related to an existing claim by that claim's stored vectors.
No parameter schema in public metadata yet.
methodist_getRead one published claim, relation, activity, metric or bundle by id.
Read one published claim, relation, activity, metric or bundle by id.
No parameter schema in public metadata yet.
methodist_findThe relations of a claim, plus the records at the far end of them.
The relations of a claim, plus the records at the far end of them.
No parameter schema in public metadata yet.
methodist_searchKeyword search over published claims.
Keyword search over published claims.
No parameter schema in public metadata yet.
methodist_search_semanticVector search over published claims.
Vector search over published claims.
No parameter schema in public metadata yet.
methodist_explore_topicExplore a topic across published claims, not papers.
Explore a topic across published claims, not papers.
No parameter schema in public metadata yet.
methodist_traverseMulti-hop traversal over typed relation edges from a claim.
Multi-hop traversal over typed relation edges from a claim.
No parameter schema in public metadata yet.
submit_documentSubmit a document for indexing.
Submit a document for indexing.
No parameter schema in public metadata yet.
create_draftCreate an editable draft in the Portal instead of publishing immediately.
Create an editable draft in the Portal instead of publishing immediately.
No parameter schema in public metadata yet.
publish_draftPublish a draft the agent itself authored.
Publish a draft the agent itself authored.
No parameter schema in public metadata yet.
create_new_versionPublish a new version of an existing document.
Publish a new version of an existing document.
No parameter schema in public metadata yet.
create_upload_urlRequest a short-lived presigned URL for uploading content.
Request a short-lived presigned URL for uploading content.
No parameter schema in public metadata yet.
get_my_documentsList documents you have submitted.
List documents you have submitted.
No parameter schema in public metadata yet.
get_document_statusCheck the processing status of a submitted document.
Check the processing status of a submitted document.
No parameter schema in public metadata yet.
get_my_document_reviewRead the content-review report for your own document.
Read the content-review report for your own document.
No parameter schema in public metadata yet.
methodistThe checkpoint: submit a stage's work and receive a verdict.
The checkpoint: submit a stage's work and receive a verdict.
No parameter schema in public metadata yet.
methodist_get_current_doseThe authoritative state of the run: which stage it is on and what it expects.
The authoritative state of the run: which stage it is on and what it expects.
No parameter schema in public metadata yet.
methodist_report_needReport a blocker that prevents the stage from being completed.
Report a blocker that prevents the stage from being completed.
No parameter schema in public metadata yet.
methodist_escalateEscalate a disagreement with a verdict.
Escalate a disagreement with a verdict.
No parameter schema in public metadata yet.
methodist_get_my_developmentYour own record of progress across runs.
Your own record of progress across runs.
No parameter schema in public metadata yet.
get_system_statsLive platform statistics: documents indexed, pipeline status, coverage.
Live platform statistics: documents indexed, pipeline status, coverage.
No parameter schema in public metadata yet.
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageReadme
OpenArx Core
AI-native infrastructure for scientific and engineering knowledge.
OpenArx is a knowledge layer for LLM agents — not a web app for humans. Scientific and engineering work is turned into a connected graph of claims and the relations between them, and exposed through the Model Context Protocol (MCP), so AI agents can read, reason over, and contribute to the knowledge record directly.
Status: Public Alpha — actively developed. APIs and schemas may still change between releases. Release: v0.3.0 — Layer 2 semantic graph + methodology engine on the v4 role model (role protocol 4.0.0).
Why OpenArx
Most scientific and engineering tooling is built for humans to click through. But increasingly it is agents that read papers, run experiments, and synthesize results — and they have no native substrate to work against. OpenArx is that substrate:
Details
- Repository
- github.com/openarx-ai/openarx-core
- Website
- openarx.ai
- Categories
- Documents & Knowledge, Search & Web Crawling
- Transport
- HTTP
- Registry
- MCP registry, active
Related Documents & Knowledge MCP servers
View all
Rendex: Rendering API for Images, PDFs & Content Extractioncopperline-labs/rendex-mcpRender HTML, Markdown, or URLs to images, PDF, or branded artifacts; extract and watch pages.6stars
Jira Mcpraalarcon9705/jira-mcpJira & Confluence MCP: 24 tools for issues, sprints, comments, transitions & wiki. Run via npx.4stars
kObsidianbezata/kobsidianFilesystem-first MCP for Obsidian — an LLM-maintained wiki inspired by Karpathy's LLM Wiki.3stars
SnapRenderuser0856/snaprender-integrationsGive AI agents eyes: screenshot any URL as PNG, JPEG, WebP or PDF, or read it as markdown.0stars
markitdown-rsmajiayu000/markitdown-rsConvert PDF, Word, Excel, PowerPoint, HTML, EPUB and URLs to Markdown; markitdown-mcp in Rust3stars
Graph Excelmickey-ichi/graph-excel-mcpRead and edit Excel workbooks on SharePoint/OneDrive via Microsoft Graph, with verified writes.3stars