Cloudflare free plan security for SaaS: what you get for $0
Free Cloudflare gives a SaaS protection others pay hundreds of dollars for. I explain what it is made of and why the order you set it up in matters.

State of knowledge: 17 November 2026.
Before you pay for any security plan, check what you already have. Free Cloudflare is not “just something to start with”. Set up properly, it gives your SaaS protection that big companies pay hundreds of dollars a month for.
Here I explain, without technical jargon, what that protection is made of and why the order you roll it out in matters more than the feature list itself.
What do you actually get for zero złoty?
Let’s start with what Cloudflare gives you on the free plan. I won’t quote the marketing. Below is what really translates into security:
| What you get | What it does for you |
|---|---|
| SSL/TLS encryption | Data between the user and the server is unreadable to anyone listening in |
| DNSSEC (DNS protection) | Nobody can redirect your users to a fake site |
| WAF (web firewall) | 5 free rules (Custom Rules). You block specific types of attacks before they reach the app |
| Rate Limiting (traffic limiter) | 1 rule. You protect login forms from automated attacks (brute-force) |
| Free Managed Ruleset | A reduced, ready set of Cloudflare rules against widely known holes. It is on every plan, the free one too |
| Global CDN | Your site is served from the server closest to the user |
| Static sites on Cloudflare Pages | Traffic to static files has no limit. The limits are 100 projects per account and 500 builds a month |
Five WAF rules and one Rate Limiting rule don’t sound like much. But every precisely configured rule blocks a whole class of attacks, and that is enough.
Plan limits change more often than you’d think. That’s why this post has a “State of knowledge” date at the top, and I checked the numbers in the Cloudflare documentation linked below the text. Why I do it this way is in the post about a “state of knowledge” date and a script that fact-checks AI articles.
Why the rollout order matters
Imagine you are building a house. You don’t start by painting the walls. First the foundations, then the structure, then the wiring and pipes. Security works the same way: each layer protects the next one.
If you set up the firewall first but forget to encrypt the connection to your server, you stop attacks at the front and leave the back door open.
Layer 1: SSL/TLS, the foundations of encryption
The first and most important decision: is the data between Cloudflare and your server encrypted?
The connection from the user to Cloudflare’s network is encrypted right away. The stretch from Cloudflare to your server depends on the mode. Today Cloudflare turns on Automatic SSL/TLS by default: it checks your server itself and picks the most secure mode that works with it, but it never lowers the setting. If the server has no valid certificate, you end up in Flexible mode (plain HTTP goes to the server) or Full (encryption without checking the certificate). Full (Strict) mode means Cloudflare verifies your server’s certificate and fully encrypts that connection (End-to-End). It is the difference between a locked safe and a shoebox with a padlock on the front.
On top of that, you set HTTP connections to be redirected to HTTPS automatically, and the minimum version of the encryption protocol to TLS 1.2 (older versions have known holes).
HSTS (HTTP Strict Transport Security): careful with subdomains
There is a setting that forces browsers to always use HTTPS on your domain. It looks great. But if you switch it on for all subdomains at once, every test environment needs a valid certificate too. Best to roll it out in steps: a short validity time first, then the full one.
Layer 2: DNSSEC, so nobody can pretend to be you
Imagine someone swaps the house number plate on your street. Your guests end up next door, with someone pretending to be you.
DNS Cache Poisoning works exactly like that. The attacker swaps your domain’s IP address in the DNS server’s answers, so users land on a fake site and type their passwords there.
DNSSEC is a cryptographic signature on every DNS answer. No server without the right key can pretend to be your domain. Switching it on in Cloudflare takes a few clicks. If your domain sits with an external registrar, you have to copy over one DS record (short for Delegation Signer).
Layer 3: WAF (Web Application Firewall), 5 precise rules
The web firewall (WAF) is the bouncer at the entrance to your app. On the free plan you have five slots for your own rules (Custom Rules). Here is how to use them best:
Rule 1: block unexpected HTTP methods and ports. Normal users talk to the server in a few standard ways (fetching pages with GET, sending forms with POST). Bots and scanners try niche methods and non-standard ports. This rule blocks everyone who is “knocking on the back door”.
Rule 2: protect authentication (Managed Challenge). Login and sign-up endpoints get attacked the most. Here Cloudflare shows an interactive challenge (Managed Challenge). Real users often pass it without even seeing it, bots get stopped.
Rule 3: block known attack vectors. Every day your app is scanned for known holes: WordPress panels (/wp-admin), configuration files (.env), database tools. None of these paths should exist in your app, so you block them in bulk, in one rule, by URI path.
Rule 4: hide staging environments. The admin panel should be reachable only from your company IP. Test subdomains (preview, stage) should not be open to the whole internet either.
Rule 5: Emergency Shield. This rule sits ready in case of a DDoS attack. When something bad is happening, one switch turns on extra verification (Managed Challenge) for all traffic to the API. Day to day it stays off, so it doesn’t slow down responses for users.
An important limit of the free plan
Free firewall rules don’t support regular expressions (advanced text patterns). The
matchesoperator only comes with the Business plan, so Pro doesn’t have it either. You can match paths by exact prefixes: starts with, ends with, contains. With a well thought out design that is enough, but it’s good to know before you configure anything.
Layer 4: Rate Limiting (a traffic limiter), one rule
Imagine someone trying to guess your password by typing a thousand combinations a minute (a brute-force attack). Rate Limiting says: “after 15 requests in 10 seconds from one IP address - stop, blocked”.
You have one rule. Point it where attacks hurt most: the login, sign-up and password reset forms.
On the free plan this rule counts requests by IP address only, in a 10-second window, and the block also lasts 10 seconds. That still slows password guessing down a lot: with a limit of 15 requests per 10 seconds, a bot makes a dozen or so tries and then has to wait.
Layer 5: Security Headers
The user’s browser can be your ally or an attack vector. It depends on what you tell it. Security Headers are a set of instructions sent to the browser with every page (for example through Transform Rules in Cloudflare).
For example: you can tell the browser that your page should never be displayed inside another page (protection against Clickjacking). Or that it may load scripts only from your domain (protection against XSS). You can set these instructions centrally in Cloudflare instead of configuring them in every app separately. On the free plan you have 10 active Transform Rules. There is also a ready switch, “Add security headers” (Managed Transforms), but it doesn’t add the Content-Security-Policy header, so you write the rule for scripts yourself.
The biggest hole nobody thinks about
You can have Cloudflare configured perfectly and still have a serious problem. A firewall won’t fix holes in the app code itself. Those are covered in the post on where vibe coding leaves holes in production, and on the portal I describe five security checks for code written with AI.
The second thing is the server. If an attacker learns your server’s IP address, they can bypass Cloudflare completely and hit the server directly.
How could they know that address? From historical DNS records (platforms archive old IP addresses), from TLS certificate logs, or by forcing your app to make an outgoing connection.
The fix is Cloudflare Tunnel (formerly Argo Tunnel): a small program (the cloudflared daemon) installed on the server, which opens the connection to Cloudflare’s network itself, from the inside. Your server needs no open incoming ports. Even if someone knows your IP address, there is nothing to connect to.
What if I use Cloudflare Pages or Workers?
Short answer: then you don’t need Cloudflare Tunnel. Cloudflare Pages and Workers are serverless services that live natively in the edge infrastructure (Edge). You have no origin server of your own with a hidden IP address or open ports, so there is no “back room” an attacker could get into past the protections. The exception: a Tunnel is useful only when your Worker has to connect to your private, external database (for example a server in the office) that you don’t want to expose publicly to the internet.
Infrastructure as code: why bother
Every configuration change made by hand in the dashboard is a change that is not in the Git history. You don’t know who made it, when, or why. During a security incident that is a serious problem.
Terraform lets you write the whole Cloudflare configuration (Infrastructure as Code) in .tf text files, version them together with the app code and deploy through CI/CD. Rolling back to the previous configuration is one command.
You don’t need it from day one. But the earlier you start, the less pain when you scale.
When does the free plan stop being enough?
Don’t start with a paid plan. Roll out the free configuration and watch for a few weeks what attacks your system and how. Only then will you know what you really need.
Signs it’s time to upgrade:
- Five WAF rules are too few and you have to choose which classes of attacks you don’t block
- You want to limit traffic separately for each logged-in user, not just by IP address (other counting characteristics only come with the Enterprise plan)
- You need the full managed rule sets: the Cloudflare Managed Ruleset and the OWASP Core Ruleset (a database of known attacks, kept up to date). Both start at the Pro plan
- You want bot analysis based on machine learning
The rule before the first dollar spent
You don’t know what you need until you measure what is attacking your app. Deploy first, then look at the data, then decide.
Common questions
Do I need to know how to code to set this up?
The basic settings (SSL, DNSSEC, simple WAF rules) can be clicked together in the Cloudflare dashboard without writing code. If you want to manage the configuration as code (Terraform), you need someone technical or an AI agent. That is what the implementation file below is for.
Does Cloudflare slow my site down?
No, it speeds it up. Cloudflare caches static assets and serves them from the network node closest to the user. The first connection is slightly slower because of the proxy layer, but every one after that is faster.
What happens if Cloudflare blocks a real user?
Cloudflare Managed Challenge (invisible verification) handles this better than a classic CAPTCHA. Real users pass automatically, bots are blocked. With well-configured rules, false alarms are rare.
Implementation file for your agent
If you have an AI agent or a developer who is going to roll out this configuration, below is a ready checklist with exact commands. It contains the complete Terraform code, verification scripts and a description of each step in rollout order.
Download the implementation checklist for your agent (Markdown)
I work with people 1:1 under the delta240mvt brand. On your project, from the first step.
See the offer ↗

