All docs

Broski

Turn a URL into article Markdown or JSON, or drive the headless browser from the command line.

Open BroskiAbout Broski

  1. Install the skim command

    Download the binary for your machine: Windows x64 (daisi-broski-skim.exe), Linux x64, macOS Intel, or macOS Apple Silicon. The command name is broski.

  2. Skim a page

    broski https://example.com/article --format md prints Markdown. --format json prints the structured article. --no-scripts skips JavaScript, which is faster when the page is already in the HTML.

  3. Call the API instead

    Sign in at Broski and create a key on the keys page. The key starts with db_. Send it as Authorization: Bearer, as X-Api-Key, or as api_key on a GET. POST https://broski.daisi.ai/api/v1/skim with a JSON body.

  4. Search and crawl

    POST /api/v1/gofer/search returns hits and does not crawl. POST /api/v1/gofer/crawl starts from seed URLs. POST /api/v1/gofer/research searches, then crawls the top unique URLs. GET /api/v1/health needs no key.

  5. Drive the browser engine

    The engine command is daisi-broski. fetch downloads and parses. run also executes inline scripts. screenshot writes a PNG and is Windows-only. By default fetch and run use a sandboxed child with a 256 MiB cap. --no-sandbox runs in-process.

Commands

broski — skim to Markdown or JSON

broski https://en.wikipedia.org/wiki/JavaScript --format md
broski https://example.com/report.pdf --format json
broski https://example.com --format md --no-scripts

daisi-broski-skim <url> [options]
  --format json|md|both     default json
  --out FILE                file instead of stdout
  --ua STRING
  --max-redirects N         default 20
  --scripts                 run scripts first (default)
  --no-scripts              skip scripts
  --quiet

HTTP skim

curl -X POST https://broski.daisi.ai/api/v1/skim \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer db_YOUR_KEY" \
  -d '{"url":"https://en.wikipedia.org/wiki/JavaScript","format":"md"}'

daisi-broski — fetch, run, screenshot

daisi-broski fetch <url> [--select <css>] [--html] [--ua <string>] [--max-redirects N] [--no-sandbox]
daisi-broski run   <url> [--select <css>] [--ua <string>] [--max-redirects N] [--quiet] [--no-sandbox]
daisi-broski screenshot <url> --out page.png [--width 1280] [--height 800]

fetch and run exit 0 on success, 1 on fetch failure, 2 if the sandbox fails, 3 on a bad command.
screenshot is Windows-only and requires --out.

Gofer

POST /api/v1/gofer/search
  { "query": "quantum error correction", "sources": "Scholarly, Community", "perProviderLimit": 10 }

POST /api/v1/gofer/crawl
  { "seeds": ["https://example.com"], "maxDepth": 0, "maxPages": 50, "stayOnHost": true }

POST /api/v1/gofer/research
  { "query": "rust programming language", "sources": "Scholarly, News", "maxCrawled": 20 }

Sources: Wikipedia, Arxiv, GitHub, HackerNews, StackExchange, CrossRef,
OpenLibrary, Reddit, DuckDuckGo, Brave, Mojeek, Gdelt, Guardian, BingNews, Grokipedia.
Bundles: Scholarly, Community, Web, News, All.
Code GraphMarkdown

Reviewed October 2026.