Legal · Policy

Agent Integrity Policy

Front End is powered by an engine that runs on our servers. Your subscription gives you and your team the right to use it. This policy explains how Front End, and any AI coding agents you use it with, are expected to treat that engine, and what you agree to when you connect.

The short version

When you connect Front End to a project, the heavy lifting happens on our hosted engine, reached securely through your access key. This policy makes sure that engine is used the way it's licensed: through the official connection, not copied, rebuilt, or worked around. For almost everyone there is nothing to do here. Connecting sets this up for you, and the rules below are handled automatically by the tool and the agents you work with.

What stays yours

Your code is yours. It lives in your repository.
Your design system is yours. The manifest (your tokens, components, and pages) is plain, declarative data you can read, edit, and version like any other file.
You can always edit what you own. Adjusting tokens and components works directly, on your own machine, with or without an active engine connection.

What runs on our engine

Some capabilities are part of the proprietary engine and run remotely as part of your subscription:

Coverage and drift detection
Promoting ad-hoc styles into your governed system
Re-grounding reused components to a project's nearest equivalents
Manifest sync and platform export

These reach the engine through the official connection. Nothing needs to be installed or hosted on your side.

What you agree to

Using Front End means agreeing to the Engine Access License, Acceptable Use Policy, and this Agent Integrity Policy, which together mean:

Licensed access, not ownership. Your subscription grants the right to use the hosted engine, not ownership of the technology behind it.
No reverse engineering or rebuilding. You won't decompile, scrape, or locally reimplement the engine's operations.
No circumvention. You won't share, forge, or work around access keys or plan limits.
A narrow non-compete. You won't use your access, or anything you learn from it, to build a competing design-governance engine.
You're responsible for your agents. If your team uses AI coding agents alongside Front End, you're responsible for keeping them within this policy. Front End configures them to follow it automatically.

How this is enforced

Because the engine runs on our infrastructure, we can see how it's accessed. Attempts to scrape, reverse engineer, or circumvent your access key are detected automatically. When we detect one, the connection is cancelled on the spot and the account is flagged, which suspends access while we review. Confirmed abuse means losing the right to use Front End.

This isn't about catching people out, it's about protecting every customer: the engine, and the governed systems built on it, stay trustworthy for everyone.

If something stops working

If your connection drops or the engine is briefly unavailable, the right move is always to reconnect or check your account, never to fake or approximate a proprietary operation locally. You can manage your key, billing, and connection at app.thefrontend.ai/connect.

And if you ever choose to stop using the hosted engine, you keep full access to the parts you own: your code and your manifest.