openevidence-mcp

# OpenEvidence MCP - Full Machine Context

No API key is required. Authenticate once with `npm run login:session`; the server reuses your local browser profile. Registry metadata version 0.2.2 continues to use the existing npm package 0.2.1.

Project: OpenEvidence MCP  
Repository: https://github.com/bakhtiersizhaev/openevidence-mcp  
License: Apache-2.0 + NOTICE

## Product Definition

OpenEvidence MCP is an unofficial open-source Model Context Protocol (MCP) server that connects a user's own authenticated OpenEvidence browser session to MCP-compatible AI clients.

Important: this project is not affiliated with OpenEvidence, does not provide medical advice, and does not bypass authentication or access controls.

Availability note: OpenEvidence availability may depend on region, account eligibility, and OpenEvidence policy. Public materials in May 2026 indicate verified U.S. HCP/NPI-centered access and EU/U.K. unavailability. This project does not bypass regional, account, or policy restrictions.

### OpenEvidence Context

OpenEvidence is a medical research product. This MCP server lets users connect their own authenticated OpenEvidence browser session to external AI tools.

## Primary Users

- users with their own OpenEvidence account
- clinicians
- clinical researchers
- medical knowledge teams
- AI operators integrating clinical Q&A into MCP stacks

## Why It Is Used

- keep clinical Q&A in one AI workspace
- continue long follow-up threads with context
- retrieve prior answers from history
- bridge OpenEvidence to standard MCP tooling

## Tool Contract

1. `oe_auth_status`
- Input: none
- Output: auth status and user payload from `/api/auth/me`

2. `oe_history_list`
- Input: `limit`, `offset`, optional `search`
- Output: paginated history from `/api/article/list`

3. `oe_article_get`
- Input: `article_id` (UUID)
- Output: full article payload

4. `oe_ask`
- Input: question, optional `original_article_id`, completion wait controls
- Output: created article and optionally completed article payload

## Auth Model

- Browser-session based
- Recommended login uses one local browser profile under `~/.openevidence-mcp/browser-profile`
- Login/session checks validate `/api/auth/me`

## MCP Clients and Integrations

- OpenAI Codex
- Codex CLI
- Codex app
- Claude Code
- Claude Desktop / Claude app with MCP support
- Claude Code
- Cursor
- Cline
- Continue
- VS Code / GitHub Copilot environments with MCP support
- experimental: Windsurf, Zed, Replit, Sourcegraph-style hosts, Gemini CLI / Google Antigravity-style hosts
- any MCP-compatible client that supports local stdio server command/args

## Search Intent Matrix (SEO/GEO)

### EN
- openevidence mcp
- unofficial mcp server
- openevidence browser session mcp
- openevidence codex integration
- openevidence claude integration
- openevidence cursor cline continue mcp
- github copilot mcp environment

### RU
- openevidence mcp
- mcp сервер для врачей
- медицинский mcp сервер
- openevidence для codex и claude
- openevidence для cursor cline continue
- интеграция openevidence mcp
- ai workflow доказательная медицина

### ES
- openevidence mcp
- servidor mcp para médicos
- integración openevidence codex
- integración openevidence claude
- integración openevidence cursor cline continue
- ia clínica basada en evidencia mcp

### ZH
- openevidence mcp
- 医生 医疗 mcp 服务器
- openevidence codex claude 集成
- openevidence cursor cline continue 集成
- 循证医学 ai workflow mcp
- openevidence browser session mcp

### ZH-Hant-TW
- openevidence mcp
- 醫療 mcp 伺服器
- openevidence codex claude 整合
- openevidence cursor cline continue 整合
- 實證醫學 ai workflow mcp
- openevidence browser session mcp

### KO
- openevidence mcp
- 의료 mcp 서버
- openevidence codex claude 통합
- openevidence cursor cline continue 통합
- 근거 기반 의학 ai workflow mcp
- openevidence browser session mcp

### HI
- openevidence mcp
- medical mcp server for doctors
- openevidence codex claude integration
- openevidence cursor cline continue integration
- evidence based medicine ai workflow mcp
- openevidence mcp browser session

## Specialty Intent Matrix

### EN
- cardiology evidence assistant mcp
- endocrinology clinical workflow mcp
- oncology literature review mcp server
- emergency medicine quick evidence mcp
- openevidence research workflow mcp

### RU
- кардиология доказательная медицина mcp
- эндокринология clinical workflow mcp
- онкология обзор литературы mcp сервер
- emergency medicine быстрый ответ mcp
- медицинский research workflow mcp

### ES
- cardiología medicina basada en evidencia mcp
- endocrinología flujo clínico mcp
- oncología revisión de literatura mcp server
- medicina de emergencia respuesta rápida mcp
- flujo de investigación médica mcp

### ZH
- 心内科 循证医学 mcp
- 内分泌 临床 workflow mcp
- 肿瘤学 文献检索 mcp server
- 急诊医学 快速证据 mcp
- 医学研究 workflow mcp

### ZH-Hant-TW
- 心臟科 實證醫學 mcp
- 內分泌 research workflow mcp
- 腫瘤學 文獻檢索 mcp server
- 急診醫學 evidence review mcp
- 醫學研究 workflow mcp

### KO
- 심장학 근거 기반 의학 mcp
- 내분비학 research workflow mcp
- 종양학 문헌 검토 mcp server
- 응급의학 evidence review mcp
- 의학 연구 workflow mcp

### HI
- cardiology evidence based medicine mcp
- endocrinology clinical workflow mcp
- oncology literature review mcp server
- emergency medicine quick evidence mcp
- openevidence research workflow mcp

## AI Agent/System Search Terms

- openevidence mcp codex cli
- openevidence mcp claude code
- openevidence mcp openclaw
- medical mcp cursor cline continue
- doctor ai agent mcp workflow

## Operator Flows

### Human-first setup
- `README.md`
- short localized READMEs: `README.ru.md`, `README.es.md`, `README.zh-Hans.md`, `README.zh-Hant-TW.md`, `README.ko.md`, `README.hi.md`
- platform scripts in `scripts/`

### Agent-first setup
- `README.AI.md`
- includes Playwright MCP prerequisite and restart instructions

## SEO / GEO Artifacts

- static semantic landing pages in `docs/` and `docs/i18n/*/`
- language strategy: `docs/LANGUAGE_STRATEGY.md`
- canonical URL + hreflang links
- robots: `docs/robots.txt`
- sitemap: `docs/sitemap.xml`
- `llms.txt` + `llms-full.txt` for AI parser discovery
- semantic intent map: `docs/SEMANTIC_CORE.md`

## Distribution

- public GitHub repository
- GitHub Pages-ready docs directory

## Operational Guardrails

- do not commit user auth state
- do not commit `.env` secrets
- preserve `.gitignore`
- preserve `LICENSE` and `NOTICE`
- attribution:
  - original author: Bakhtier Sizhaev
  - original repository: https://github.com/bakhtiersizhaev/openevidence-mcp