http_getLABSCO SUMMARY
Requests that get through, and a reader for what comes back.
Ordinary HTTP tools get turned away because their TLS handshake does not look like a browser's; this one matches the fingerprint down to JA3/JA4 and HTTP/2. The half that makes it useful daily is the second one — a long page or a PDF comes back as Markdown, or as a line range you can page through.
What it is
A full HTTP client built on rquest, with accurate TLS, JA3/JA4 and HTTP/2 browser fingerprints, plus response storage and PDF and HTML to Markdown conversion.
What you get
- All eight HTTP methods: GET, POST, PUT, DELETE, PATCH, HEAD, OPTIONS, TRACE
- Browser fingerprints at the TLS, JA3/JA4 and HTTP/2 layers, over BoringSSL
- Headers, cookies, redirects, query parameters, form data, JSON payloads and multipart uploads
- Basic, Bearer and custom authentication
- Large responses stored in a system temporary directory and read back by line range
- HTML and PDF responses converted to Markdown for the model to read
- The PDF model loading state readable, and restartable if it stalls
Requirements
Python, installed from PyPI as mcp-rquest. PDF to Markdown conversion uses the Marker library, whose models load in the background — get_model_state reports where that has got to.
Setup effort
One command — pip install mcp-rquest
USE CASES
01
Reading a page that blocks scripted requests
The request carries a real browser fingerprint, and the response comes back converted to Markdown rather than as raw HTML.
02
Turning a PDF you found into readable text
Fetch it, then convert; if the conversion models are still loading, the state call tells you rather than the request simply failing.
03
Testing an API you are building
Every write method is available with JSON, form or multipart bodies and the auth scheme you use.
04
Checking what a server says without pulling the body
Headers and allowed methods answered on their own, which is usually all a redirect or CORS question needs.
05
Working through a response too big for one turn
The full body is stored on disk and read back by line range, so nothing is truncated in transit.
06
Getting PDF conversion going again after it stalls
Restart the model loading and watch the state rather than restarting the whole server.
12 tools
Listed in project docs. Your client may show a different list.
http_posthttp_puthttp_deletehttp_patchhttp_headhttp_optionshttp_traceget_stored_responseget_stored_response_with_markdownget_model_staterestart_model_loadingBEFORE YOU INSTALL
What to know first
Tools you get
12 tools named in the project's documentation.
Requires
uvx on PATH
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
Server setup
TERMINAL
pip install mcp-rquestMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"http-rquest": {
"command": "uvx",
"args": [
"mcp-rquest"
]
}
}
}
