Skip to main content
Glama
JondaX MCP Server

JondaX MCP Server

Official
README.md
# JondaX Model Context Protocol (MCP) Server

Official Model Context Protocol (MCP) server for **JondaX by Jonda Health**. Connect AI assistants (Antigravity, Claude Desktop, Cursor, Zed, Cline) directly to the JondaX Health Data Transformation Engine.

JondaX extracts data from clinical files — PDFs, images, HL7, FHIR — normalises it, and returns clean, structured, coded output in the format your system needs.

---

## Capabilities & MCP Tools

JondaX provides two core processing modules and retrieval tools:

| Tool Name | Type | Description |
| :--- | :--- | :--- |
| **`upload_pathology_scan`** | **Async** | Uploads a pathology report or lab document (PDF, JPG, PNG, JSON, CSV, Parquet and etc). JondaX de-identifies, extracts, normalises, translates, and returns an `uploadId` for tracking. |
| **`upload_medical_device`** | **Sync** | Uploads an image of a medical device reading (pulse oximeter, blood pressure monitor, glucometer, etc.) and extracts readings instantly. |
| **`get_upload_status`** | **Query** | Checks processing status (`uploaded`, `processing`, `completed`, `failed`) and webhook delivery status. |
| **`list_uploads`** | **Query** | Lists your recent uploads (newest first) with `uploadId`, file name, status, upload date and expiry. Optional filters: `limit` (1–100, default 10), `fromDate`/`toDate` range, and `status`. Useful for discovering `uploadId`s. |
| **`get_extracted_results`** | **Query** | Retrieves structured extracted results (JSON, FHIR JSON, FHIR XML, HL7, CSV, Parquet). Retained for 30 days. Returns HTTP 202 if still in progress. |

---

## Configuration

Get your API key from the **Integration** section in the JondaX Client Portal. https://app.jondax.eu

| Environment Variable | Required | Default | Description |
| :--- | :--- | :--- | :--- |
| `JONDAX_API_KEY` | **Yes** | — | Your JondaX API token (`jondax_...`) |
| `JONDAX_BASE_URL` | No | `https://app.jondax.eu` | JondaX API Base URL |

---

## Quickstart Setup in AI Tools

### 1. In Claude Desktop (`claude_desktop_config.json`)

```json
{
  "mcpServers": {
    "jondax": {
      "command": "npx",
      "args": ["-y", "@jondax/mcp-server"],
      "env": {
        "JONDAX_API_KEY": "YOUR_JONDAX_API_KEY",
        "JONDAX_BASE_URL": "https://app.jondax.eu"
      }
    }
  }
}
```

### 2. In Cursor / Antigravity (`mcp_config.json`)

```json
{
  "mcpServers": {
    "jondax": {
      "command": "npx",
      "args": ["-y", "@jondax/mcp-server"],
      "env": {
        "JONDAX_API_KEY": "YOUR_JONDAX_API_KEY",
        "JONDAX_BASE_URL": "https://app.jondax.eu"
      }
    }
  }
}
```

---

## How It Works

```
┌──────────────┐     upload_pathology_scan      ┌─────────────────────────────┐
│              ├───────────────────────────────►│ JondaX Transformation Engine│
│   AI Agent   │                                │  • De-identifies & Extracts │
│ (Antigravity │◄───────────────────────────────┤  • Maps Medical Terminology │
│   / Claude)  │       Returns uploadId         │  • Converts Units & Codes   │
│              │                                └──────────────┬──────────────┘
│              │     get_extracted_results                     │
│              ├───────────────────────────────────────────────┘
│              │◄───────────────────────────────
│              │   Returns Structured Results
└──────────────┘   (JSON, FHIR, HL7, CSV, etc.)
```

---

## Supported Output Formats & Standards

* **FHIR**: FHIR JSON, FHIR XML (Observation, DiagnosticReport)
* **HL7**: HL7 v2
* **Standard Data**: JSON, CSV, Parquet
* **Terminology & Coding**: LOINC, standard clinical units & normalized ranges

---

## Compliance & Security

JondaX is built specifically for healthcare data:
* **HIPAA** · **GDPR** · **PDPA** · **ISO 27001**

---

## Local Development

If contributing or building from source:

```bash
git clone https://github.com/JondaHealthTech/jondax-mcp-server.git
cd jondax-mcp-server
npm install
npm run build
```

---

## Documentation & Links

* **API Docs**: [jondax.redocly.app](https://jondax.redocly.app/)
* **Product**: [jonda.health/product/jondax](https://www.jonda.health/product/jondax)
* **Client Portal**: [app.jondax.eu](https://app.jondax.eu)

## License

MIT © [Jonda Health](https://www.jonda.health)

TDQS

A4/5.0

Scored across 4 tools

Disambiguation4/5

The two upload tools are distinguished by document type (pathology/blood test vs medical device), and the two retrieval tools are clearly separated by status vs results. There is minor potential confusion because both upload tools trigger the same OCR/extraction pipeline, but descriptions are clear enough.

Naming Consistency5/5

All tool names follow a consistent verb_noun pattern: upload_* for ingestion and get_* for retrieval. The naming is predictable and easy to navigate.

Tool Count5/5

Four tools is well-scoped for the server's purpose: upload two categories of medical documents, check status, and retrieve results. Each tool earns its place with no unnecessary redundancy.

Completeness4/5

The core workflow of upload → status → results is fully covered. Minor gaps exist around error details or listing past uploads, but agents can successfully complete the primary extraction workflow without dead ends.

Maintenance

ActivityStale
ResponsivenessNo recent community issues