securityproduct

How Eldrin Keeps Your Work Safe

When you hand an AI the keys to build software for you, a natural question follows: how do I know my stuff is safe?

It's a fair question. Most AI-powered tools send your code to a remote server, process it somewhere you can't see, and hope you trust them with whatever you're building. Eldrin takes a fundamentally different approach.

Here's how — in plain language, no jargon.

Everything Stays on Your Computer

This is the single most important thing to understand about Eldrin: your projects never leave your machine.

There is no cloud server storing your files. No remote database holding your ideas. No sync service copying your work somewhere else. When you create a project in Eldrin, it lives in a folder on your computer — right next to your photos and documents.

The only time anything goes over the internet is when the AI needs to think. Eldrin sends a question to the AI model and gets an answer back. That's it. Your project files, your notes, your plans — they all stay right where they are.

Your Secrets Are Locked in a Vault

To use Eldrin, you need an API key — think of it as a password that lets the AI work on your behalf. Storing passwords as plain text is a terrible idea, so Eldrin doesn't do that.

Instead, your API key is encrypted using the same security system your computer already trusts. On a Mac, that's the Keychain — the same place Safari stores your website passwords. On Windows, it's the Credential Manager. These are hardened systems built by Apple and Microsoft specifically to protect sensitive information.

Even if someone copied the file where the key is stored, they'd see nothing but scrambled nonsense without access to your computer's security system.

Every Project Gets Its Own Space

Imagine each of your projects as a separate room in a building. Eldrin makes sure the doors between rooms are locked.

Each project lives in its own folder with its own files, its own history, and its own database entry. The AI agents working on one project cannot see or touch another project. They don't even know the other rooms exist.

This means if you're building a personal budget tracker and a side project for a friend, the two are completely separate. One can never accidentally affect the other.

The AI Works in a Sandbox

Here's where it gets interesting. When Eldrin's AI agents write code for you, they don't run loose on your computer. They work inside a sandbox — a controlled space with clear boundaries.

Think of it like hiring a contractor to renovate your kitchen. You give them access to the kitchen, but you don't hand over the keys to your entire house. They can rearrange cabinets and install a new sink, but they can't wander into your bedroom or go through your filing cabinet.

In Eldrin, this works through several layers:

The workspace boundary. Agents can only see and modify files inside your project folder. They can't access your desktop, your downloads, or your other projects.

Protected infrastructure. Certain files in your project are locked — the AI can't modify them. These are the structural files that keep everything running smoothly. The AI can build on top of the foundation, but it can't dig up the foundation itself.

Network restrictions. When running in protected mode, agents can only connect to the internet for specific, expected purposes — downloading software packages, accessing the AI model, or pushing code to GitHub if you ask them to. They can't make random connections to unknown servers.

Optional: A Fortress Around the AI

For users who want the strongest possible protection, Eldrin can run the AI agents inside what we call a Fortress — a completely isolated environment that's separate from your main computer.

Think of it as a sealed room within a room. The AI does all its work inside this protected space, and even if something went wrong, nothing could escape to affect your actual computer.

On a Mac, this is a lightweight virtual machine. On Windows, it uses a built-in isolation technology called WSL2. Either way, the AI agent sees only the project it's working on and nothing else.

The Fortress also has its own firewall — a set of rules that controls what the AI can and can't do on the network. By default, it blocks everything except the connections it needs to do its job.

A Complete History of Every Change

Every change the AI makes to your project is recorded. Eldrin uses the same version control system that professional software teams rely on — Git.

This means you can see exactly what was changed, when it was changed, and why. If an AI agent adds a new feature or fixes a bug, there's a record. If something doesn't look right, you can go back to any previous version with a single action.

Nothing is hidden. Nothing is overwritten without a trace.

No Hidden Connections

Some apps quietly send usage data, crash reports, or analytics to their makers. Eldrin doesn't.

There are no tracking scripts. No usage analytics. No telemetry. The app explicitly disables these systems. The only network traffic comes from actions you initiate — asking the AI to think, downloading a package, or deploying your project when you're ready.

You can verify this yourself. Every connection Eldrin makes is visible and auditable.

What This All Means for You

You don't need to understand encryption algorithms or firewall rules to use Eldrin safely. The protections are there by design — built into the architecture, not bolted on as an afterthought.

Here's the short version:

  • Your files stay on your computer. Always.
  • Your credentials are encrypted. Using the same system that protects your other passwords.
  • Your projects are isolated. From each other and from the rest of your system.
  • The AI is contained. It can only do what it needs to do, in the space you give it.
  • Every change is recorded. You can always see what happened and go back if needed.
  • No surprises. No hidden data collection, no silent uploads, no mystery connections.

Building software with AI should feel empowering, not risky. That's what we're here for.