# OpenArx

OpenArx gives AI agents MCP access to scientific papers, evidence checks, and a connected claim graph for research work.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 8
- Author: OpenArx-AI
- License: Apache-2.0 · policy=link-only · description only; the file stays with the author
- Source: https://github.com/OpenArx-AI/openarx-core
- Updated: 2026-07-30
- HTML: https://affpapa.org/en/nevibecode/mcp/openarx

## What it does

OpenArx connects an MCP-compatible agent to scientific and engineering knowledge. It can search papers, retrieve document details and claim data, compare publications, inspect research methods, and locate related work. Search supports both semantic matching and exact keyword lookup.

Claims and their links form a graph stored alongside a vector index, with Neo4j used for graph storage. Relations can record support, extension, qualification, or refutation. Agents may also submit claims and relations through a staged methodology review that filters unsupported contributions.

A bearer API token from `portal.openarx.ai` is required, and the researcher endpoint is `https://mcp.openarx.ai/researcher/mcp`. OpenArx is in Public Alpha, so its APIs and schemas can change between releases.

## Requirements
- Remote server: `https://mcp.openarx.ai/researcher/mcp`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http openarx https://mcp.openarx.ai/researcher/mcp
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add openarx --url https://mcp.openarx.ai/researcher/mcp
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "openarx": {
            "url": "https://mcp.openarx.ai/researcher/mcp"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "openarx": {
            "type": "http",
            "url": "https://mcp.openarx.ai/researcher/mcp"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http openarx https://mcp.openarx.ai/researcher/mcp
```

## When to use it
- search scientific papers by topic or exact term
- check a claim against published evidence
- compare several papers side by side
- trace links between related claims

## Example prompts
- Search the corpus for studies on this method and group the supporting evidence.
- Check this claim against the papers, separating support from contradiction.
- Compare these papers by research method and reported findings.

## Similar
- [nature-reviewer](https://affpapa.org/en/nevibecode/skills/nature-reviewer.md) — skill · 13900 installs
- [literature-search-openalex](https://affpapa.org/en/nevibecode/skills/literature-search-openalex.md) — skill · 3100 installs
- [literature-search-arxiv](https://affpapa.org/en/nevibecode/skills/literature-search-arxiv.md) — skill · 3400 installs
- [mirofish-offline-simulation](https://affpapa.org/en/nevibecode/skills/mirofish-offline-simulation.md) — skill · 1400 installs
- [neo4j-document-import-skill](https://affpapa.org/en/nevibecode/skills/neo4j-document-import-skill.md) — skill · 923 installs
- [neo4j-agent-memory-skill](https://affpapa.org/en/nevibecode/skills/neo4j-agent-memory-skill.md) — skill · 879 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md