Vellum Kasane

Turn company names or websites into their public job board: which ATS (Greenhouse, Lever, Ashby, Workable, SmartRecruiters), the board slug and URL, open-job count and confidence. Reads the careers page and verifies against the vendors' public APIs. Output plugs into ATS Jobs Feed.

Job listings from the public JSON APIs of Greenhouse, Lever, Ashby, Workable and SmartRecruiters. One clean record per job (Markdown description, department, locations, remote flag, salary when published, apply URL) plus a diff mode that flags new, updated and closed jobs between runs.

Turn any documentation site into a clean Markdown corpus: one file per page with front matter (title, URL, breadcrumbs), headings, code blocks with languages, tables and absolute links, plus an INDEX.json. Built for LLM wikis, RAG and offline docs. Pay per page.

Export issues, pull requests, discussions and releases from public GitHub repos as clean records or as Markdown wiki notes with YAML front matter. Optional comments, label/state/date filters, token support for higher rate limits. Pay per item.

Hacker News exporter using the official Firebase and Algolia APIs: front/new/best/ask/show/jobs lists, search, full story threads as Markdown, and Who-is-Hiring threads parsed into structured jobs (company, role, location, remote, salary hints). Pay per story or job.

Export Substack, Ghost and beehiiv newsletter archives as clean Markdown with YAML front matter: title, authors, date, tags, paywall flag, word count. Built for LLM wikis, RAG and research. Only public content; paywalled posts return the free preview. Pay per post.

Search PubMed and Europe PMC, or pass PMIDs, PMCIDs and DOIs, and get LLM-ready Markdown with YAML front matter: abstract, authors, MeSH, keywords, citations, license. Open-access papers include full text with sections, tables and references. Official API only. Pay per article.

Fetch RSS, Atom and JSON feeds, then extract each entry's full article as clean Markdown (title, author, published, content, word count, content hash). Dedupes across runs with a named store. Ideal for LLM wikis, newsletters and monitoring. Pay per entry.