Webfold API

Web, research and bookmark endpoints over HTTP: scrape, crawl, map and search web pages, search and retrieve arXiv papers, fetch YouTube transcripts, archive pages as Markdown, convert documents, read PDFs with OCR, redact personal data, build corpora, check translations, and run sandboxed computers.

Authenticate every request except GET /health with a bearer API key. Fetched content is cached; the x-cache response header reports hit, miss or raw-hit. Errors have a plain-text body, not JSON. The hosted API is in early access.

Web

Fetch pages as Markdown, crawl a site, discover its URLs, and search the pages you have crawled.

• POST
  Scrape a web page/v1/web/scrape
• POST
  Crawl a web site/v1/web/crawl
• POST
  Discover URLs on a web site/v1/web/map
• POST
  Search the local web index/v1/web/search

Research

Search arXiv, retrieve papers as chunked Markdown, and fetch YouTube transcripts.

• POST
  Search arXiv/v1/arxiv/search
• POST
  Retrieve an arXiv paper/v1/arxiv/retrieve
• POST
  Fetch a YouTube transcript/v1/youtube/transcript

Bookmarks

Archive a page as Markdown at creation time, tag it and search your archive.

• POST
  Create and archive a bookmark/v1/bookmarks
• GET
  List and search bookmarks/v1/bookmarks
• GET
  Get a bookmark/v1/bookmarks/{id}
• PATCH
  Update a bookmark/v1/bookmarks/{id}
• DELETE
  Delete a bookmark/v1/bookmarks/{id}

Collections

Group bookmarks. Deleting a collection deletes its bookmarks unless unlink=true.

• POST
  Create a collection/v1/collections
• GET
  List collections/v1/collections
• GET
  Get a collection with its bookmarks/v1/collections/{id}
• DELETE
  Delete a collection/v1/collections/{id}

Account

Per-meter usage for your key.

• GET
  Usage report/v1/usage

System

Service health. No authentication.

• GET
  Health check/health

Documents

Convert office documents, e-books and spreadsheets to Markdown.

• POST
  Convert a document to Markdown/v1/documents/convert

OCR

Read PDFs, including scans.

• POST
  Extract text from a PDF, with OCR for scans/v1/ocr

PII

Detect and redact personal data.

• POST
  Redact personal data from text/v1/pii/redact
• POST
  Check text for personal data/v1/pii/scan

Corpus

Redact, chunk and de-duplicate text into training or retrieval corpora.

• POST
  Redact a batch of records/v1/corpus/redact
• POST
  Chunk and de-duplicate records into corpus rows/v1/corpus/build

Localization

Extract translatable strings and verify translations.

• POST
  Verify translations against their sources/v1/localization/check
• POST
  List the translatable strings in a file/v1/localization/extract

Computers

Sandboxed computers an agent can drive, with human takeover and an audit trail.

• POST
  Create a sandbox computer/v1/computers
• GET
  Get a computer/v1/computers/{id}
• DELETE
  Destroy a computer/v1/computers/{id}
• GET
  List capability tokens/v1/computers/{id}/tokens
• POST
  Mint a capability token/v1/computers/{id}/tokens
• POST
  Revoke every token/v1/computers/{id}/tokens/revoke-all
• DELETE
  Revoke one token/v1/computers/{id}/tokens/{token_id}
• POST
  Resume the agent/v1/computers/{id}/agent/resume
• GET
  Agent guard state/v1/computers/{id}/agent/guard
• POST
  Take over from the agent/v1/computers/{id}/takeover
• POST
  Hand control back to the agent/v1/computers/{id}/handback
• POST
  Revoke agent access/v1/computers/{id}/revoke
• POST
  Suspend a computer/v1/computers/{id}/suspend
• POST
  Unsuspend a computer/v1/computers/{id}/unsuspend
• GET
  Read the audit log/v1/computers/{id}/audit
• GET
  Where to open the live desktop/v1/computers/{id}/stream