Menu

Routelint

Find what your APIs leak.

Record the calls a vibe-coded app already makes. Replay them with no session and as a peer. Read redacted evidence—no OpenAPI required.

Captured call

Critical

POST /api/search

{ "email": "[REDACTED]", "invoiceId": "inv-9001" }

Anonymous · 200 · same body returned to a peer account

What every security scan checks

Five test suites →

The product

API security testing from live UI traffic.

A browser session records the calls your vibe-coded app already makes. Those calls are replayed with no session and as a peer account. Forgotten paths are probed from a short list.

Primary · API security

Replay the calls the interface already made.

No OpenAPI required. Capture XHR and fetch from the running app, then send each call again as anonymous, User A, and a peer. Findings cover unguarded routes, IDOR, secrets, and sensitive paths.

  • Prove the host with a DNS TXT record or a well-known file.
  • One scan without an account shows counts and severities.
  • An account opens the five suites, identities, and redacted evidence.
  • A login scan can send the bearer token you paste.

Also available · SEO & GEO

A free public check for search and answer engines.

Paste a URL for scores on robots.txt, sitemap, llms.txt, on-page signals, and crawler access. Separate from the API suites—it does not scan your APIs.

SEO check

What the public check is looking for.

Search is scored from technical, on-page, and on-site signals. Answer engines are scored from crawler access, raw text, structure, and dates.

SEO & GEO overview

Technical

robots.txt, the sitemap, HTTPS, redirects, a viewport tag, and how long the page takes to answer.

On-page

Title length, the meta description, a single h1, image alt text, and whether the page has enough words.

On-site

A link to About, a link to Contact, an author, and a date a reader can find on the page.

Crawlers

GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and CCBot, plus a short /llms.txt map of the site.

Without JavaScript

Sentences that are already in the HTML, so a crawler that never runs a script can still read the page.

Structure and dates

Headings, lists, and JSON-LD for the organization, an article, a product, or a person, with a modified date.

The product

A site that runs the suites for you.

Point Routelint at an application you own. It watches the interface, keeps the API surface, and checks whether those routes stay inside the right account.

01

Capture

Playwright follows the live UI and records XHR and fetch traffic, including query strings and JSON bodies.

02

Replay

Each captured call is sent again as anonymous, User A, and a linked peer, with the original URL and body.

03

Report

Findings land in one place: suite, severity, route, and redacted evidence.

Features

Five checks on every scan.

Each suite answers one question: what the UI called, what returns without login, what a peer can read, what secrets appear, and which forgotten paths still answer.

Guides

Short answers before the scan.

Each page answers a question people search before they know the name Routelint, then points at a redacted example or the public check.

All guides

How it works

A scan stays inside the host you name.

Replay stays read-only. Evidence is redacted before it reaches the database.

How it works
  1. 01

    Add the host

    Paste the app URL, then prove the host with a DNS TXT record or a well-known file. The scan stays read-only.

  2. 02

    Record the UI

    A browser session follows the running app and keeps each XHR and fetch: method, query, body, status, and response.

  3. 03

    Replay identities

    The same call is sent again with no session, as User A, and as a peer account.

  4. 04

    Store redacted evidence

    Tokens, emails, cards, and password hashes are stripped before a finding is written.

An account

Signup opens the rows the public pages hold back.

The five suites are included on an individual account. Payments and daily scans are coming soon.

Normal scan

The crawler opens the app with no session. It records the calls a signed-out visitor can already make.

Login scan

Paste a bearer token on a target you added. That scan sends Authorization: Bearer, so pages behind login are in reach. The token is stored encrypted.

Full SEO report

The public check shows the scores and the name of each row. An account adds the evidence behind it.

Three sites

A free account holds three targets. Identities, captured calls, and redacted findings belong to that account.

Questions

What a scan includes.

A scan without an account shows counts. Signup opens the suites. Payments and daily scans are coming soon.

What is Routelint?
Routelint is an API security scanner that records the APIs a running vibe-coded app already calls, then replays them with no session and as a peer account to find unguarded routes, IDOR, secrets, and forgotten paths. A separate free SEO and GEO check scores how a public site reads to search and answer engines.
How do I do API testing for a vibe-coded app?
Start from the live UI, not an OpenAPI file. Routelint records the XHR and fetch calls your vibe-coded app already makes, then replays each call with no session and as a peer account to find unguarded data leaks and IDOR. Use the API leak scan or a free site scan on a host you control.
Can I test APIs without an OpenAPI specification?
Yes. Routelint treats the running interface as the spec. A browser session captures method, URL, body shape, and status from real traffic, then the suites replay those observed calls. You do not upload a Swagger or OpenAPI document to start.
What is IDOR and how does Routelint find it?
IDOR (insecure direct object reference) is when User B can read User A’s object by reusing the same URL or body. Routelint captures a call as User A, then replays it with a peer identity. Matching account data in the peer response is the finding.
Can Routelint check a Lovable, Cursor, or v0 app?
Yes, if you control the host. Prove ownership, then Routelint records browser API traffic from the running app and runs the same five suites. Many AI-built apps ship UI-first without OpenAPI; that is the workflow Routelint is built for.
Do I have to prove I own the host?
A security scan does. Publish the token as a DNS TXT record, or at /.well-known/routelint-verification.txt, and Routelint checks it before the scan runs. The public SEO check does not. Production targets still ask for an extra confirmation.
What does a scan without an account show?
One scan per browser returns counts and severities. Titles, routes, and evidence stay on the server until you sign in and claim the scan.
What does an account include?
Signup is an individual account. It includes the five test suites, the full security report, identities, and the SEO and GEO check rows. Payments and daily scans are coming soon.
What does the SEO and GEO check cover?
It reads public HTML: the page, a few links on the same host, robots.txt, an XML sitemap, and llms.txt. Scores, row names, and pass, warn, or fail are public. The evidence behind a row stays on the server until you sign in. The security suites stay on the scan product.

Run API security testing on a vibe-coded app you control.