Burp Suite
Burp Suite is the industry-standard platform for testing web applications. It sits between your browser and the server, letting you see, pause, modify and replay every request — which is how almost every web vulnerability is actually found.
What it is
Burp Suite is an intercepting proxy plus a toolbox built around it. A proxy is simply a middleman: your browser sends the request to Burp, Burp shows it to you, and then forwards it to the server. Nothing about the traffic changes unless you change it — but now you can.
That single capability is why Burp matters. A browser only lets you submit the requests the page designer intended. Burp lets you submit any request: change a price, swap a user ID, remove an authorisation header, inject a payload into a JSON field, replay a request 500 times with different values. Almost every serious web vulnerability — IDOR, SQL injection, SSRF, authentication bypass, business-logic flaws — is found by manipulating requests a browser would never send.
Two editions exist. Community is free and includes the proxy, Repeater, Decoder, Comparer and a limited Intruder. Professional is paid and adds the automated scanner, unlimited Intruder speed and project saving. Everything you need to learn web security is in the free edition — the PortSwigger Web Security Academy labs are built around it.
How to use Burp Suite
Follow it exactly as you would in a real session — every command, every line of output, and what each result actually tells you. This is a interface walkthrough.
Set the scope before you touch anything
Browse the app with interception OFF, then read history
/api/ endpoints, and those endpoints often check authentication but forget authorisation.Send it to Repeater and change ONE value
Authorization: Bearer eyJ…
"customer":"another.user@target.local",
"amount":18450.00}
Automate only what repeats — Intruder
Confirm with Comparer when the difference is subtle
Flags & options
| FLAG | WHAT IT DOES |
|---|---|
| Proxy → Intercept | Pause requests as they pass. For surgery, not browsing. |
| HTTP history → Filter | Show only in-scope traffic so noise disappears. |
| Ctrl+R / Send to Repeater | Manual edit and resend — where most findings happen. |
| Ctrl+I / Send to Intruder | Automated payloads at marked positions. |
| Sniper / Cluster bomb | Attack types: one payload set, or one per position. |
| Decoder | Base64, URL, hex, JWT — read opaque tokens. |
| Comparer | Diff two responses; spot the one field that changed. |
| Target → Scope | Restrict capture and testing to what you are authorised for. |
| Proxy → Options → TLS | Per-host TLS settings when a target uses a client certificate. |
Why it matters
A browser can only send the requests the developer intended. Almost every serious web bug — IDOR, broken access control, business-logic abuse — exists in the gap between what the page allows and what the server actually checks. Burp is how you stand in that gap: replay the same request with one value changed and watch the server's answer.
How it works
The core loop is: intercept → inspect → modify → forward → observe. You configure your browser to send traffic through Burp (default listener 127.0.0.1:8080), install Burp's CA certificate so HTTPS works, and turn on interception when you want to pause a request.
In practice you rarely intercept every request — that makes browsing unusable. The professional workflow is the opposite: leave interception off, browse the application normally so Burp records everything in the HTTP history, then right-click any interesting request and send it to Repeater to modify and resend it manually, over and over, watching how the response changes.
The tools map to tasks:
- Proxy — capture and modify live traffic.
- Repeater — manual request editing and replay. This is where you will spend most of your time.
- Intruder — automated requests with payload positions; used for fuzzing parameters, enumerating IDs, testing wordlists. Community edition is rate-limited.
- Decoder — encode/decode Base64, URL, hex, JWT pieces.
- Comparer — diff two responses to spot subtle differences (a classic way to find blind injection).
- Scanner (Pro) — automated crawling and vulnerability checks. Useful, but it finds what it is programmed to find; manual testing finds logic flaws.
Sits between browser and server. Every request and response passes through and is logged in HTTP history.
USE IT FORMapping the application. Browse logged-in and logged-out; the API calls in history are usually where the real logic lives.
GOTCHAIntercept-on is for surgery, not for browsing. Leave it off and work from history.
Take any request from history, edit it, resend, and compare responses side by side.
USE IT FOREvery hypothesis you have. Change one value, send, read the diff. This is where 80% of real findings are made.
GOTCHAChange one thing at a time. Two changes at once means you cannot tell which one mattered.
Automates the same request with different payloads at marked positions (Sniper, Battering ram, Pitchfork, Cluster bomb).
USE IT FOREnumeration once you know a parameter matters: user IDs, coupon codes, filenames, password lists.
GOTCHACommunity edition is rate-limited by design. Sort results by length or status — the outlier is the finding.
Decoder handles Base64, URL, hex, HTML and JWT segments. Comparer diffs two responses at word or byte level.
USE IT FORReading opaque tokens, and spotting the one field that changed between two otherwise identical responses — the classic way to confirm blind injection.
GOTCHANever trust a client-side "encrypted" cookie — decode it first.
Commands & syntax
# 1. Proxy → Options → confirm the listener 127.0.0.1:8080 # 2. In your browser (or FoxyProxy), point HTTP+HTTPS at 127.0.0.1:8080 # 3. Visit the CA cert while proxying, download and install it http://burp/cert # 4. Set scope so you only capture the target Target → Scope → Add → https://target.local
# Browse the app normally with interception OFF. # Then filter history by: Proxy → HTTP history → Filter → Show only in-scope items # Right-click a request → Send to Repeater (Ctrl+R)
# Change one thing at a time and compare responses GET /api/user/1042 HTTP/1.1 # try 1041, 1043 → IDOR? # Remove the auth header entirely Authorization: Bearer ... # delete → 200 means broken access control # Watch status code, length and body — not just the rendered page
# Send request to Intruder, mark the payload position GET /product?item=§1001§ HTTP/1.1 # Attack type: Sniper · Payload: numbers 1000-1100 · Start attack # Sort by Length or Status — outliers are your findings
# Decoder: paste a JWT / cookie, decode Base64, URL-encode payloads # Comparer: paste two responses → Words diff → spot the one field that changed # Repeater tabs: keep one tab per vulnerability hypothesis
kali@errordex:~$ cat request.txt # captured in Burp, edited in Repeater GET /api/v1/invoices/1042 HTTP/1.1 Host: portal.target.local Authorization: Bearer eyJhbGciOiJIUzI1NiIs... Accept: application/json --- change ONE thing: 1042 → 1041 --- GET /api/v1/invoices/1041 HTTP/1.1 --- response --- HTTP/1.1 200 OK Content-Type: application/json { "invoice": 1041, "customer": "another.user@target.local", "amount": 18450.00, "status": "paid" } # 200 with someone else's data = IDOR (broken object level authorisation) # One number changed. That is the entire bug class.
Common mistakes
Testing without written permission
Web testing against a live site you do not own is illegal, full stop.
Intercept on all the time
You will drown. Use history + Repeater instead.
Not setting scope
You end up testing third-party services the app calls, which you have no authorisation for.
Trusting the automated scanner
Scanners miss business-logic flaws entirely — the flaws that pay the most in bug bounties.
Ignoring response headers and cookies
Missing HttpOnly, weak SameSite, permissive CORS — these are findings too.
Forgetting Burp changes nothing by itself
It is a window into traffic; the thinking is yours.
Practice safely
The best free lab in existence: PortSwigger Web Security Academy — hundreds of labs, from "what is SQL injection" to expert-level business logic, all built to be solved with Burp Community. It is free and it is the same company that makes Burp.
Work through the Apprentice labs in order: SQL injection → authentication → access control → SSRF → business logic. That sequence mirrors how real assessments unfold.