Skip to content
dreamitbuildit.ai

What we do to keep your projects safe

This is what runs on our platform today. Where something is still being rolled out, we say so.

  • Each project in its own space

    On a shared server, every project runs sealed off from the others. It can't see their files, programs or databases, and it can't connect to their services.

  • Security updates every day

    Our servers install the operating system's security fixes by themselves every day. We also update the platform's own software and the AI tools on every server, and a notice tells you while yours is being updated.

  • Firewall rules on every server

    Each project can reach only what it needs. Attempts to reach other projects, our private network or our internal services are blocked, counted and shown to our team.

  • A security sensor

    We're adding a sensor to our servers that watches for behaviour typical of an attack, such as programs started from temporary folders or software that mines cryptocurrency, and alerts our team. We're turning it on server by server.

  • Limits on sending email

    Every mailbox can send a set number of messages a day (100 for a project's mailboxes unless you ask for more), so a site that's broken or misused can't flood people with spam from our servers.

  • Encrypted backups

    Your project's backups are encrypted before they leave its server. Our control panel is backed up every 4 hours, encrypted, to two different cloud providers.

  • Passkeys for our team

    Everyone on our team with admin access signs in with a passkey, which can't be guessed or typed into a fake sign-in page the way a password can. You can add one to your own account in Settings.

  • Secure connections

    Every site we host is served over https, so what your visitors type and see travels encrypted.

Why it never stops

Software is written by people, and people make mistakes. Every day, dozens of new weaknesses are added to the public list of known vulnerabilities, in programs that millions of sites depend on. See the list (opens in a new tab)

Once a weakness is public, attackers race to use it before sites are fixed. Staying safe means installing fixes quickly, every time. That's our job, not yours.

Real cases, explained simply

Some well-known cases: what happened, and what each one teaches.

  1. 2026

    PHP2Shell

    CVE-2026-109618, CVE-2026-109619

    In October 2026 the makers of PHP, the language behind a large share of the world's websites, fixed two flaws that let an attacker run their own commands on a web server without signing in. Official advice said every PHP installation that serves web pages was affected. Five days after the fix, CISA, the US cybersecurity agency, listed both as being used in real attacks.

    Sites made with Dream It Build It run on Node.js, not PHP. It still shows how little time there is: attacks started within days of the fix.

    Source:CISA alert (opens in a new tab)UK NCSC notice (opens in a new tab)

  2. 2024

    The xz backdoor

    CVE-2024-3094

    For more than two years, someone gained the trust of the volunteers behind xz, a common compression tool, then slipped in hidden code that could have let them into servers. An engineer noticed that logins had become half a second slower, looked into it, and found it before it reached most systems.

    This is a supply-chain attack: the danger came in through a trusted update. It was caught because someone looked closely at something odd, which is what watching for unusual behaviour is for.

    Source:CISA alert (opens in a new tab)

  3. 2024

    regreSSHion

    CVE-2024-6387

    Researchers at Qualys found that a bug fixed back in 2006 had crept back into OpenSSH, the program used to log in to most Linux servers. Under the right conditions it could give an attacker full control of a server without a password.

    Old mistakes can come back. Updates have to keep coming, even for software that has been safe for years.

    Source:Qualys (opens in a new tab)

  4. 2021

    Log4Shell

    CVE-2021-44228

    Log4j is a small piece of code that a huge number of Java programs use to write down what they're doing. In December 2021 it turned out that one specially written line of text, typed into a chat box or a search field, could make it download and run an attacker's code.

    A tiny part nobody thinks about can put whole systems at risk. Knowing exactly what software you run is half the job.

    Source:CISA guidance (opens in a new tab)

  5. 2014

    Heartbleed

    CVE-2014-0160

    A mistake in OpenSSL, the software behind the padlock on a large part of the web, let anyone ask a server for a little of its memory, again and again. That memory could hold passwords and the server's secret keys.

    Installing the fix wasn't enough: sites also had to replace their keys and ask people to change their passwords.

    Source:heartbleed.com (opens in a new tab)

Security words, in plain language

Vulnerability
A weakness in software that someone could use to do something they shouldn't, like reading private data or taking control.
Exploit
The method or code that takes advantage of a vulnerability. If the vulnerability is an unlocked window, the exploit is climbing through it.
CVE
Short for Common Vulnerabilities and Exposures: the public list that gives each known vulnerability a number, like CVE-2021-44228, so everyone means the same thing.
Zero-day
A vulnerability attackers use before the makers of the software know about it, so they had zero days to prepare a fix.
Patch
A fix for a vulnerability, delivered as an update. Installing it closes the window.
Malware
Any software made to do harm: steal data, spy, damage files or use a computer without permission.
Ransomware
Malware that locks or scrambles your files and demands money to give them back. Recent backups kept somewhere else are the best defence.
Phishing
A fake message, often an email or text that looks like it's from a company you know, that tries to get you to type a password, pay something or open a file.
Privilege escalation
When an attacker who got in with limited access finds a way to get more, up to full control of the computer.
Supply-chain attack
An attack on something many people rely on, like a piece of shared code or an update, so the harm reaches everyone who uses it.
Cryptojacking
Secretly using someone else's computer or server to mine cryptocurrency. It slows everything down and runs up the bill.
DDoS
Short for distributed denial of service: flooding a site with so much fake traffic from many computers that real visitors can't get in.
Encryption
Scrambling data so only someone with the right key can read it.
Firewall
Rules that decide which connections may go in and out of a computer or a network.
Passkey
A way to sign in with your phone or computer instead of a password. It only works on the real site, so a fake page can't steal it.

What you can do

  • Use a passkey, or a different strong password for every account. A password manager makes this easy.
  • Be wary of messages that rush you. When in doubt, open the site yourself instead of clicking the link.
  • Keep your phone, computer and browser up to date.
  • Check whether your email address was in a known data breach on Have I Been Pwned, and change those passwords.
  • If something on your site looks wrong, tell us from the Support tab in the app.

Where to learn more