The web, torn open. The signal, distilled.

ReamKira turns websites into source material your AI can use. Firefox for your AI. Give it a public URL and receive readable text, headings, links, image references, metadata and structured-data summaries, with capture limits clearly reported. Copy the result into a chat or download Markdown or JSON.

Your AI can call it directly. The free GET and POST API lets agents and applications retrieve the same package without using this page, inspect what was captured, and compare content hashes between visits. No account or API key required. ReamKira reads page source; it does not execute JavaScript or inspect image contents. Client-rendered apps and content that loads after interaction may come back incomplete.

REAMKIRA // DISTILLATION TERMINAL · v0.8

AI-first web extraction

Feed the machine
something useful.

Paste a public web address. ReamKira will return the page's readable text, structure, metadata and useful links as one portable Markdown package.

Public HTTP and HTTPS pages only. No accounts, logins or private network addresses. Submitted URLs are fetched by our server (hosted in Singapore; ReamKira keeps no database and saves no submitted URLs or reports) — see Terms & Privacy for details.

package.ready

function distill_web()

Most websites are built for eyes, fingers and advertising systems. ReamKira converts the same page into the parts an LLM works best with: explicit metadata, semantic headings, readable text and resolved source links.

01 // FETCHRetrieve the public source page.

02 // REAMRemove scripts, styling, menus and repeated noise.

03 // PACKAGEReturn a clean Markdown payload ready to paste or upload.

TRANSMISSION PROTOCOL

API & URL usage

Give agents and applications a structured view of a public page: content, schema summaries, language declarations, indexing signals, links, capture quality and a text hash for repeat visits.

URL arguments, JSON API & limits

URL ARGUMENTS

Open ReamKira with the source already loaded. Add autostart=1 to begin reaming immediately.

https://www.voidneon.com/reamkira/?url=https%3A%2F%2Fexample.com%2Farticle&autostart=1

urlThe encoded public HTTP or HTTPS page address.

autostart=1Start automatically after loading the page.

content=0|1Include readable page content. Default: 1.

style=0|1Include extracted style information. Default: 1.

structure=0|1Include semantic page structure. Default: 1.

links=0|1Include resolved links. Default: 1.

images=0|1Include image references. Default: 1.

excludeComments=0|1Truncate content at the first detected comment-section marker. Default: 0 (comments included).

JSON API

POST JSON to the distillation endpoint. Read package for structured fields, markdown for ready-to-use text, and stats for size/counts. Full parsed JSON-LD remains in package.structuredData.jsonLd.

https://www.voidneon.com/reamkira/distill.php

{"url":"https://example.com/article","includeContent":true,"includeStyle":true,"includeStructure":true,"includeLinks":true,"includeImages":true,"excludeComments":false}

Method: POST · Content-Type: application/json
GET also works: distill.php?url=example.com. No login or API key required. GET and POST return identical fields.

API LIMITS

Published so integrators can plan around them. Even generous limits are better than undefined ones.

curl -X POST "https://www.voidneon.com/reamkira/distill.php" -H "Content-Type: application/json" -d '{"url":"https://example.com/article"}'

Rate limit20 requests per 60 seconds per IP address. Exceeding it returns HTTP 429.

Max URL length4,096 characters.

Max POST body20,000 bytes (20 KB).

Max page fetch size2.5 MB per source page; up to 350 KB per linked stylesheet.

Timeout15s to connect, 20s total per fetch. Slow sources return an error rather than hanging.

RedirectsUp to 6 followed before giving up.

Output capsMain text capped at 160,000 characters; up to 160 headings, 350 links and 160 images per export.

CORSNot enabled. Call the endpoint from server-side code (curl, a backend script) or GET/POST it directly — not from client-side JavaScript on another domain.

ErrorsAlways JSON: {"error":"message"}, with an appropriate HTTP status (400 invalid input, 405 wrong method, 413 payload too large, 422 fetch/processing failed, 429 rate limited).

LoggingReamKira itself does not log requests or store submitted URLs. Our hosting provider may keep standard server access logs for security and operations — see Terms & Privacy.

RELEASE TRANSMISSION

Current version · v0.8

Updated 17 September 2026 · Melbourne, Victoria

Richer schema summaries, canonical and language declarations, page signals and link counts. Built for clearer decisions by humans and AI.

Version history & update log

Download updates.txt

Open this panel to load the dated update log.

Feed the machinesomething useful.