API Reference

Integrate FHE computations into your backend via REST API. All endpoints are accessed through your API key.

Quick Start

All M2M requests use a single authentication header:

x-api-key: sk_live_...

Generate an API key from Dashboard > API Keys. Your company is identified automatically from the key.

Test it yourself, no partner needed

You don't need a second company to try the full flow. Create an internal permission (your company granting to itself) with grantType: "internal". It skips the two-party key handshake, so there's no counterparty to wait on: encrypt your own data with the toolkit, upload it, execute, and decrypt the result yourself (internal results finish atsucceeded, with no partial-decryption release step).

Everything cryptographic happens in the offline toolkit on your own machine. Your keys and plaintext never reach JuLenny. So this self-test is also how you verify, first-hand, that we only ever hold ciphertext.

Connect an AI agent (MCP)

The JuLenny MCP serveris a small program that runs on your machine and lets an AI agent (Claude Desktop, Cursor, and other MCP clients) drive the platform on your behalf. It calls this same REST API with your key over the standard Model Context Protocol, so nothing new is exposed, and the agent can only do what your key's permissions allow.

Use “Edit Config”, not “Add custom connector”.

This is a localMCP server, so add it through your client's config file, in Claude Desktop: Settings → Developer → Edit Config. The “Add custom connector” dialog is for remote, hostedMCP servers authenticated with OAuth; it expects a server URL and does not accept a static API key, so it won't work for this server.

1. Get an MCP key

An administrator enables MCP for your user (Dashboard > Users), then you generate a dedicated MCP key under Dashboard > API Keys. The key is tied to your account and inherits your permissions (read-only or read-write).

2. Add the server to your MCP client config

Paste this into your client's MCP config (e.g. Claude Desktop'sclaude_desktop_config.json), replacing the key with your own:

{
  "mcpServers": {
    "julenny": {
      "command": "npx",
      "args": ["-y", "@julenny/mcp-server"],
      "env": {
        "JULENNY_API_KEY": "sk_live_...",
        "JULENNY_API_URL": "https://julenny.net"
      }
    }
  }
}

3. Environment variables

JULENNY_API_KEY: your MCP key (required).
JULENNY_API_URL: the platform base URL: https://julenny.net

Restart your MCP client after editing the config. The agent then has tools to list collaborations, upload datasets, trigger executions, and fetch results, each gated by your key's permission level, so a read-only key can browse but not run or modify.

Endpoint Reference

Complete list of REST API endpoints for FHE operations. Click an endpoint to view request/response examples.

API Reference

Quick Start

Test it yourself, no partner needed

Connect an AI agent (MCP)

1. Get an MCP key

2. Add the server to your MCP client config

3. Environment variables

Endpoint Reference

Authentication

Datasets

Functions

Collaborations

Permissions & Grants

Key Setup

Execution