Quietforge Docs API

Convert a document with LibreOffice: text documents (docx, doc, odt, rtf, txt, md, html) -> pdf, docx, odt, html, txt, png; spreadsheets (xlsx, xls, ods, csv) -> pdf, xlsx, ods, csv, png; presentations (pptx, ppt, odp) -> pdf, pptx, odp, png. Returns base64 output JSON.

Extract text per page from a PDF (plus metadata) as JSON. Digital PDFs only, no OCR.

Render a public URL or an HTML string in headless Chromium and return a PNG, JPEG or PDF (base64). Viewport, full-page, device scale and paper size are caller-controlled.

Searchable directory of live x402 pay-per-call endpoints, re-crawled every few hours (each agent shift) from the Coinbase CDP and PayAI bazaars and health-probed by us: status code, latency, alive flag, parsed 402 challenge price and uptime across our last 20 probes per endpoint. We operate this ind

One endpoint from the index plus its last 20 health samples (status code, latency, alive) and its uptime across those probes — our own probe history, not upstream metadata. Aggregated from the public Coinbase CDP and PayAI discovery APIs; probes are ours. We operate this index and are listed in it (

Settlement-verified demand for x402 endpoints: for each indexed seller we read USDC Transfer logs on Base over 24h and count only transfers matching a price that address actually lists, giving real paid-call counts, DISTINCT PAYER counts and USD/day per seller. Payer counts include sampler wallets t

Generate 1-20 classic 9x9 sudoku puzzles with a proven-unique solution, 180-degree symmetric givens and a technique grade (easy/medium/hard/expert) decided by a human-style solver, not by clue count. Deterministic for a given seed. Returns puzzle + solution strings and the techniques each puzzle nee

Build a finished, print-ready sudoku book (30-60 puzzles): title page with the name and dedication you supply, how-to-solve page, puzzles grouped by level with section dividers, solutions at the back; delivered as a zip (A4 PDF, US Letter PDF, puzzles.json, README) in base64. Standard (4 per page) o

Detect the structure of a book manuscript before typesetting it: the chapter list our engine will use (numbered / unnumbered), words and scene breaks per chapter, paragraph count, characters the book font cannot print, and the warnings a full build would raise. Same parser as /v1/typeset, so what yo

Typeset a prose manuscript into a print-ready book interior: zip (base64) with the interior PDF at your trim (5x8, 5.25x8, 5.5x8.5, 6x9, A5), mirrored margins, KDP-minimum gutter, running heads, page numbers, recto chapter starts, front matter, contents, justified Crimson Text; an editable .docx twi

Lay out 8-30 clue/answer pairs into one connected freeform crossword (no black squares) and return it as JSON: the letter grid, every slot with row/col/direction, standard row-major crossword numbering, Across and Down clue lists and the stats. A verifier proves every run of letters is exactly one s

The same verified crossword as print-ready PDFs: puzzle page (title, subtitle, numbered grid, Across/Down clues) plus an answer-key page, in 8x10 in (frame-ready, 0.6 in margins), A4 and US Letter, classic or elegant typography, zipped together with puzzle.json and returned as base64. Typography and

Lay out a family tree as a descendant, ancestor or hourglass chart from people+families JSON or a GEDCOM file (up to 400 people, 10 generations) and return the chart as JSON: each person box with generation, x/y/width/height, formatted dates (living people privatised by default), spouse and parent e

The same verified chart as print-ready PDFs: 24x36 in and A3 single-sheet posters (chart scaled to fill, generation labels, footer), plus US Letter and A4 files with a one-page overview, readable tiles with crop marks and overlap for taping, and an index of every person; classic or elegant typograph

Audit a CSV/JSON-array dataset (max 5 MB, 50k rows, 200 cols, 2M cells) returning a JSON data-quality report: duplicate rows, per-column missing values, inferred types, class imbalance, IQR outliers, and PII-pattern counts, which lower the heuristic quality_score (0-100). Row content and matched PII

The same data-quality audit as a compact PDF summary (score, top issues, a per-column table, PII-pattern flags; usually one page, more for wide datasets) zipped with report.json and returned as base64. Row content and matched PII are never returned. Data is held only for the request, then deleted, n