17 Services from workbot1.oddsys.org

Fuzzy-matches up to 50 names against the US OFAC SDN list, returning scored matches with associated sanctions programs and aliases.

Generates 384-dimensional L2-normalized mean-pooled sentence embeddings using all-MiniLM-L6-v2 for up to 96 texts per call

Fetches a public web page and returns clean markdown or structured JSON, reducing token count by 80-85% compared to raw HTML.

Returns live ECB daily foreign-exchange reference rates for any currency pair, with optional amount conversion

Crop, resize, convert to greyscale, and re-encode images (PNG/JPEG/WebP) in a single deterministic API call

Detects and locates objects in an image using YOLOX-nano (COCO classes), returning per-class counts and pixel bounding boxes for 80+ object categories.

Validates and normalizes phone numbers using Google libphonenumber, returning validity, multiple format representations, country, region, and line type for single numbers or batches up to 100.

Generates a QR code from input text, returning it as a base64-encoded PNG or SVG with customizable size, margin, and colors

Reranks up to 64 candidate texts against a query using a cross-encoder model, returning ranked indices with 0–1 relevance scores for precision-stage RAG pipelines.

Decodes a Base-mainnet transaction hash or raw EVM calldata into human-readable form — function selector, matched signature, named arguments, and ERC-20/USDC transfer events — and can verify an x402 payment settlement receipt in one call.

Converts structured data between JSON, YAML, CSV, and XML formats with deterministic, edge-case-handled output.

Converts raw HTML into clean, token-efficient markdown, typically reducing token count by 80-85% before passing content to an LLM.

Extracts clean markdown text from text-layer PDFs without OCR, converting unreadable documents into compact, LLM-friendly tokens.

OCR a scanned PDF (up to 25 pages) into clean text using Tesseract for English machine-printed documents; text-layer PDFs are extracted directly.

Splits a document into token-bounded chunks at natural boundaries (paragraphs/sentences) with optional overlap, producing ready-to-embed pieces for RAG pipelines.

Extracts structured JSON from a web page including title, byline, publish date, Open Graph/meta tags, JSON-LD schema.org data, links, and images — deterministically, with no LLM.

Performs OCR on scanned PDFs up to 100 pages using Tesseract (English, machine-printed), with a 240-second time budget and graceful partial-result truncation.