Skip to content
MCP ThesaurusMCP Thesaurus

URLs LE

CommunityIncomplete39/100Claim

npm urls-le-mcpstdioupdated 24d ago

Pull every URL out of the current file in one keystroke Any text file β€” Markdown, HTML, CSS, JavaScript, TypeScript, JSON, YAML, Properties, TOML, INI and XML know what to exclude

SourceWebsite

What can you do with URLs LE?


Useful? A star or rating is how other developers find it β€” β˜… GitHub Β· β˜… Open VSX Β· β˜… Marketplace

What it does

Open a file, press Ctrl+Alt+U (Cmd+Alt+U on Mac), and every URL in the document lands in a new editor β€” deduplicate and sort it from there. Works in VS Code and in VS Code–based editors like Cursor and VSCodium (installable from Open VSX).

  • Link auditing β€” every link, autolink, and plain URL in Markdown and HTML (code blocks and comments excluded)
  • Source review β€” URLs in string literals, template literals, and comments across JS/TS
  • Config sweep β€” URLs in JSON strings, YAML values, Java properties, TOML/INI values, and XML (Maven POMs, feeds)

Install

Where What you get Install
VS Code The extraction, in your editor, on a keystroke Marketplace
Cursor, VSCodium, Windsurf The same extension Open VSX
A terminal or a CI step The same run over a whole tree, with exit codes cargo install urls-le Β· crates.io
Any MCP agent, via Node extract_urls over stdio npx urls-le-mcp Β· npm
Zed The MCP server as a context server zed-industries/extensions#7077 (pending review)

Use it from an AI agent

The same extraction engine runs as an MCP server, so an agent can call it directly instead of you running a command.

Editor How
VS Code 1.101+ Nothing to install β€” the extension registers extract_urls with agent mode
Zed URLs-LE β€” pending review
Claude Code claude mcp add urls-le -- npx -y urls-le-mcp
Cursor, Windsurf, anything else point it at npx urls-le-mcp
extract_urls(content, format?, filename?, dedupe?, maxResults?)

Returns every URL with its protocol and 1-based line and column, capped at 500 by default with meta.truncated so a large file cannot flood the agent's context window.

The server takes content and returns data β€” it reads no files and makes no network requests of its own. Published as urls-le-mcp on npm and as io.github.nolindnaidoo/urls-le in the MCP registry.

Most hosts read a JSON config. Add one entry:

{
  "mcpServers": {
    "urls-le": {
      "command": "npx",
      "args": ["-y", "urls-le-mcp"]
    }
  }
}

-y skips the install prompt on first run. Pin a version if you would rather not track releases β€” urls-le-mcp@2.3.1.

Prefer not to go through npx on every launch? Install it once and point at the binary instead:

npm install -g urls-le-mcp
{
  "mcpServers": {
    "urls-le": { "command": "urls-le-mcp" }
  }
}

It speaks MCP over stdio and needs no environment variables, no API key and no configuration of its own. To check it before wiring it into anything:

echo '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' | npx -y urls-le-mcp

That prints the tool list and exits β€” if you see extract_urls, the server works.

The CLI

The same extraction runs from a terminal or a shell pipeline: a Rust CLI in crate/, sharing one corpus with the extension β€” crate/fixtures/ β€” so the two can never read a document differently.

urls-le .                     # every URL in the tree
urls-le --dedupe docs/        # one line per distinct URL
urls-le mcp                   # the same extraction over MCP on stdio

# the point of the whole thing:
urls-le . | jq -r '.urls[].value' | sort -u | lychee -

Exit codes follow grep β€” 0 URLs found, 1 none found, 2 the question was malformed β€” so if urls-le src/; then … works and finding nothing is an answer rather than an error.

It has no opinions, deliberately. No link checking, no insecure-scheme flag, no filtering. An http:// URL is wrong in a production config and right in a test fixture; a tool that decides for you is one you configure, then argue with, then mute β€” and the muting takes the extraction with it. Pipe it to lychee and let that have the opinions.

Supported formats

Format Language IDs Notes
Markdown markdown Fenced code blocks and inline code excluded
HTML html <!-- --> comments excluded (multi-line supported)
CSS css Quoted or bare url(...), @import
JavaScript / TypeScript javascript, typescript Strings, template literals, comments
JSON json String literals only, via jsonc-parser token offsets; comments are trivia
YAML yaml Whole content, comments included
Properties properties #/! comment lines excluded
TOML toml Parsed values only; comments excluded
INI ini Whole content; ;/# comment lines excluded
XML xml Attributes and text content
Anything else any Whole document scanned; fileType reports unknown

No document is refused. The eleven above know what to exclude; every other language id β€” python, go, shellscript, csv, plaintext, log, whatever your editor calls it β€” is scanned whole, because a URL is unambiguous in any text and there is nothing in those worth excluding.

Extracted protocols: http, https, ftp, file, mailto (requires an @), tel. Every occurrence is reported with its real line and column β€” TOML positions are forward-located in the source and can be approximate for repeated identical values. A URL ends at whitespace or a quote/bracket delimiter, so relative links (/docs) and bare domains (example.com) are never extracted, and URLs containing raw spaces extract as space-terminated partials. Trailing ./, are kept β€” they are legal URL characters.

Commands

Command Description
URLs-LE: Extract URLs (Ctrl+Alt+U / Cmd+Alt+U) Extract all URLs from the active document
URLs-LE: Deduplicate URLs Remove duplicate lines from the results
URLs-LE: Sort URLs Sort results alphabetically, by domain, or by length
URLs-LE: Open Settings Open URLs-LE settings
URLs-LE: Help Built-in documentation

Settings

Setting Default Description
urls-le.openResultsSideBySide true Open results beside the current editor
urls-le.postProcess.openInNewFile true Open results in a new file (when not side-by-side)
urls-le.copyToClipboardEnabled false Also copy results to the clipboard
urls-le.dedupeEnabled false Deduplicate extraction results automatically
urls-le.notificationsLevel silent all = every notification, important = warnings + errors, silent = errors only
urls-le.safety.enabled true Guardrails for very large files
urls-le.safety.fileSizeWarnBytes 1000000 Refuse extraction above this file size
urls-le.safety.largeOutputLinesThreshold 50000 Warn above this line count
urls-le.statusBar.enabled true Show the status bar item
urls-le.telemetryEnabled false Local-only event log (see Privacy)

Languages

Twelve languages besides English:

German Β· Spanish Β· French Β· Indonesian Β· Italian Β· Japanese Β· Korean Β· Portuguese (Brazil) Β· Russian Β· Ukrainian Β· Vietnamese Β· Chinese (Simplified)

Both halves are covered β€” the manifest (command titles, setting names and descriptions) and everything shown while the extension runs (notifications, the status bar, the sort quick-pick). The extension follows VS Code's display language, so it matches whatever the editor is already set to; no setting of its own.

Privacy & security

  • No network access. The extension never fetches, validates, or pings the URLs it extracts β€” it only reads the text of the active document. The telemetryEnabled setting writes events to a local Output Channel you can inspect (URLs-LE Telemetry); nothing leaves your machine.
  • The MCP server holds the same line. It takes content as an argument and returns data: no filesystem access, no network calls, no telemetry. Your agent already has file-read tools, so duplicating them inside the server would add a path-traversal surface for no capability. check:mcp-bundle fails the build if the server ever imports something that could reach either.
  • Error notifications redact home directories and credential-shaped fragments.

Documentation

What Where
What the tool is allowed to say β€” scope, output contract, refusals, non-goals crate/SPEC.md
How the extension is built and held together β€” architecture, invariants, toolchain, release AGENTS.md
How the CLI is built and held together crate/AGENTS.md
What changed CHANGELOG.md Β· crate/CHANGELOG.md
The tool's page, and the other fifteen letools.dev/tools/urls-le

Performance

Input Size Found Time Rate Scan speed
Markdown docs 2.63 MB 50,000 61.14 ms 817,802/sec 43 MB/s
HTML page 1.25 MB 30,000 21.88 ms 1,370,972/sec 57 MB/s
JSON config 1.52 MB 40,000 39.01 ms 1,025,358/sec 38.8 MB/s

Median of 7 runs after warmup, on Apple M5 Pro, 24 GB RAM, Node 24.3.0. Inputs are generated by scripts/benchmark.ts rather than checked in, so the sizes above are exactly what was measured. Reproduce with bun run benchmark.

These are machine-specific and are not asserted in CI β€” a benchmark that gates a build only tells you how busy the runner was.

Testing

Metric Coverage
Statements 94.18%
Branches 84.48%
Functions 94.33%
Lines 94.44%

322 test cases across 22 files, plus an integration suite that runs in a real VS Code extension host and an end-to-end test that installs the built .vsix into a clean profile.

Generated from a real run β€” coverage/coverage-summary.json and coverage/test-results.json β€” by scripts/coverage-readme.js; CI fails if this section drifts. Reproduce with bun run test:coverage, and the case count is the one vitest prints.

More from the LE family

Sixteen single-purpose tools for the work in front of every model. Each ships a Rust CLI and an MCP server. One page: letools.dev

Get it out

  • String-LE β€” Extract every string in a codebase, with its position, so a person can read them
  • Numbers-LE β€” Extract every hardcoded number in a codebase, so a person can check them
  • Units-LE β€” Extract every quantity with its unit, normalized, and refuse the ambiguous ones by name
  • Dates-LE β€” Extract every date and timestamp, and the exact instant each one resolves to
  • IDs-LE β€” Extract every UUID, ULID, NanoID, ObjectId and Snowflake, and decode the time inside
  • IPs-LE β€” Extract every IP address, CIDR block and MAC, normalized and classified by scope
  • URLs-LE β€” Extract every URL in a codebase, with its protocol and exact position
  • Paths-LE β€” Extract every file path in a codebase, and say whether it still points at anything
  • Colors-LE β€” Extract every color in a codebase, and say which ones are not in your palette

Check it

  • Regex-LE β€” Find every regex in a codebase, and report which can be driven into catastrophic backtracking
  • Versions-LE β€” Find where one dependency is constrained differently across a repository's manifests
  • i18n-LE β€” Identify the i18n library a project uses, then audit its catalogs by that library's rules
  • Scrape-LE β€” Check whether a page is scrapeable before the scraper is written, and say when it cannot tell

Guard it

  • Secrets-LE β€” Find hardcoded credentials in a codebase, and never print one into the report
  • EnvSync-LE β€” Compare the dotenv files in a tree, and say which keys are missing from which
  • Unicode-LE β€” Find the Unicode that hides meaning β€” bidi controls, invisibles, homoglyphs, mixed scripts

Each stands on its own: no shared crate, no published core. Where two of them agree, it is because the same answer was right twice.

Contact β€” nolindnaidoo.com Β· GitHub Β· LinkedIn

Also by nolindnaidoo

Rust β€” pixelcoords and pixelactions are one loop: pixelcoords answers where, pixelactions acts there. Their own tools, their own voice β€” not part of the LE family.

License

MIT Β© nolindnaidoo