Free browser tool · zero cloud lock-in
External Link Status Checker
Audit reference URLs from notes for SSRF, private-network exposure, redirects, and broken endpoints.
Runs 100% locally in your browser. Audit reference links from your Markdown notes for private network exposure, SSRF risks, and invalid ports. Generates safe offline curl audit scripts and parses HTTP verification status logs.
Step 1: URL Security & SSRF Audit
Paste URLs or Markdown text. Links are parsed and audited against private-network and cloud metadata security rules.
| Status | Destination URL | Classification | Security Warning |
|---|---|---|---|
| Safe Public | https://developer.apple.com/documentation/screencapturekit | safe-public | None |
| Safe Public | https://commonmark.org/help/ | safe-public | None |
| Safe Public | http://github.com/marqly/oyma-web | safe-public | None |
| Safe Public | https://broken-reference-sample.org/old-spec.pdf | safe-public | None |
| Safe Public | https://anti-bot-protected-site.com/research-paper | safe-public | None |
| Flagged | http://127.0.0.1:8080/internal-metrics | loopback | Local loopback destination: requests never reach public web |
| Flagged | http://localhost:3000/dev-preview | loopback | Local loopback destination: requests never reach public web |
| Flagged | http://192.168.1.100/router-admin | private-network | Private LAN / Intranet RFC 1918 IP address (192.168.1.100) |
| Flagged | http://10.0.0.5:9000/internal-db | private-network | Private LAN / Intranet RFC 1918 IP address (10.0.0.5) |
| Flagged | http://169.254.169.254/latest/meta-data/iam | cloud-metadata | CRITICAL SSRF TARGET: AWS/GCP/Azure link-local instance metadata endpoint |
| Flagged | https://internal.corp.network:6379/redis-dump | suspicious-port | Unusual internal service/database port (6379) |
Step 2: HTTP Response Log Parser & Dashboard
Paste the output from your curl verification script or server access logs to categorize HTTP status codes.
| Code | Category | Target URL | Effective Destination |
|---|---|---|---|
| 200 | 200 OK (Live) | https://developer.apple.com/documentation/screencapturekit | — |
| 200 | 200 OK (Live) | https://commonmark.org/help/ | — |
| 301 | 301 Redirected | http://github.com/marqly/oyma-web | https://github.com/marqly/oyma-web |
| 404 | 404 Dead / Not Found | https://broken-reference-sample.org/old-spec.pdf | — |
| 403 | 403 Bot-Blocked / Forbidden | https://anti-bot-protected-site.com/research-paper | [Cloudflare Bot Block] |
| 500 | 500 Server Error | https://crashed-server-sample.net/api/v1 | — |
| 0 | Timeout / Error | https://timeout-domain-sample.invalid/data | [Connection Timed Out] |
The Problem with Direct In-Browser Link Checking
Many note-taking browser extensions attempt to test links by issuing asynchronous fetch() or Image() requests directly from the client. In practice, this introduces critical technical flaws:
- Cross-Origin Restrictions (CORS): Modern web browsers strictly block web scripts from inspecting HTTP response headers or status codes of third-party domains unless explicitly permitted by CORS headers.
- Server-Side Request Forgery (SSRF) in Local Context: If a note contains URLs referencing local intranet endpoints (such as
http://192.168.1.1or cloud metadata169.254.169.254), automated browser requests can trigger accidental state changes on your router, local databases, or local container daemon. - False 403 Bot Blocks: Automated headless HTTP requests lacking complete browser TLS fingerprints are routinely rejected with
403 Forbiddenby anti-scraping platforms like Cloudflare and Akamai, tricking users into believing active links are broken.
Deterministic Offline Verification Workflow
This utility resolves these problems through an isolated, two-step verification workflow:
- Security Boundary Audit: Your URLs are audited locally for SSRF risks, loopback addresses, and suspicious database ports. Unsafe targets are isolated before any requests are formulated.
- Safe Command Generation: The tool generates an offline Bash/curl script equipped with conservative connect timeouts (
--connect-timeout 8), total maximum duration limits (--max-time 15), redirect traversal limits (--max-redirs 5), and a standard macOS browser User-Agent header. - Log Categorization: When you execute the script in your local terminal and paste the resulting log back into Step 2, status codes are classified into Live (200), Redirected (301/302), Broken (404/410), and Bot-Blocked (403/Cloudflare).
Related Vault Health & Markdown Tools
Explore companion offline utilities for maintaining link hygiene and vault integrity:
- Broken-Wikilink Finder: Audit internal missing notes and unresolved double-bracket wikilinks.
- Safe URL Cleaner for Notes: Strip invasive UTM tracking parameters from note links while keeping destinations intact.
- Markdown-to-Wikilink Converter: Convert external CommonMark links to native Obsidian wikilinks without modifying web URLs.
- Vault Health Inspector: Check vault-wide link density, orphaned notes, and attachment hygiene.
- All Browser Tools: Full collection of zero-network Markdown note tools.
Frequently Asked Questions
Questions
Why does this tool avoid making raw HTTP requests directly from the browser?
Browser cross-origin security (CORS) blocks direct HTTP probes to third-party servers. More critically, initiating automatic background requests to arbitrary URLs from your browser risks server-side request forgery (SSRF), leaks your IP address, or hits intranet admin dashboards. Our tool audits URL safety in local memory and generates a constrained curl script for you to run locally.
What private-network targets and SSRF risks does this checker detect?
The checker identifies RFC 1918 private IPv4 subnets (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16), local loopback hosts (127.0.0.1, localhost, ::1), cloud metadata endpoints (169.254.169.254), and unusual database/management ports (22, 3306, 5432, 6379). These destinations are flagged as unsafe for public fetch.
How does the generated curl verification script handle timeouts and redirects?
The generated script includes --connect-timeout 8, --max-time 15, and --max-redirs 5, preventing hanging requests on dead servers or infinite redirect loops. It also uses a standard browser User-Agent header to avoid false-positive 403 blocks from bot scrapers.
How do I interpret HTTP 403 or bot-blocked responses?
Many modern platforms (Cloudflare, academic paywalls, enterprise documentation) block command-line bots and return HTTP 403 Forbidden even though the page is live for human visitors. Our status log parser labels these separately as Bot-Blocked rather than Dead Links, avoiding false alarm removals.
Are my notes or URLs sent to any remote server or third-party service?
No. The entire classification engine runs 100% client-side in browser memory. No URLs, document contents, or audit results are ever sent across the network.
Keep your notes private, portable, and fast
Write your daily records in clean, future-proof plain Markdown on your local disk with zero cloud lock-in.