Turn a hard public web page into rows you can keep working.

Careers boards, PDF bid notices, license registries: the pages a scraper comes back empty from.

One request. That is it.

Paste the page URL or describe the list you need. We take it from there.

Rows

Source URL: Every row links to the page it came from.
Checked date: You see when each row was last read.
No guessing: What could not be verified is marked, not filled in.

Three rows from profiles.health.ny.gov · directory hospitals
Name	Checked
A.O. Fox Memorial Hospital	2026-10-12
Adirondack Medical Center-Lake Placid Site	2026-10-12
Albany Medical Center - South Clinical Campus	2026-10-12

Rows on the page219

Read

2026-10-12

Source

profiles.health.ny.gov · directory hospitals

What comes back

One of three answers comes back, right here.

Already available

We already keep this page in our data sources: you get its rows now and can have them refreshed weekly.

profiles.health.ny.gov · directory hospitalsread 2026-10-12

Three rows from profiles.health.ny.gov · directory hospitals
Name	Checked
A.O. Fox Memorial Hospital	2026-10-12
Adirondack Medical Center-Lake Placid Site	2026-10-12
Albany Medical Center - South Clinical Campus	2026-10-12

Not yet

We log your request. It can then go on the bounty board as a shared bounty, where anyone can add money toward it. Your words go with it.

On the bounty board nowSee all

Cover www.federalreserve.gov: title, link, pubDate, descriptionwaiting for example answers
Cover www.federalreserve.gov: title, url, formatwaiting for example answers
美国FDA新食品添加剂监管合规实施方案（中文详版）Pays $270
Find a wrong file that passes "Standard checks: a table"Pays $23

Or put a price on yours and start today

Cannot

We say why, plainly. Public http(s) pages only.

Cannot extractthe reason, plainly

Why: the address is not a public page we can reach
What next: nothing is guessed and no request is opened, so send a different page

The pages that break scrapers

Real captures of pages a scraper comes back empty from. Each is kept working; where it is in our data sources, the card links to its rows.

OpenAI careers
captured 2026-07-30 · One page of JavaScript. 303,462 characters. The listings are tangled inside a script blob, not in HTML you can parse.
```
window.__appData = {"ddRumApplicationId":"80e0bf43-…","environment":"production",…
```
303,462 bytes163 lines
Town of Hempstead bid notices
captured 2026-07-30 · 2,295 lines of HTML. 4 bid notices buried in them.
```
__VIEWSTATE" value="tCgaRlbJSGjs0UnkJZnqq96zZwnriJanm2bYaxU3nJzksq+/DS7lPjBHUnStev66e…
```
94,376 bytes2,295 lines

Village of Bayville bid notices

captured 2026-07-30 · 626 lines of markup. The notices are scattered link tags.

<a href="…/Bid-Notice-Surplus-Equipment-2008-Ford-Ranger.pdf" title="Bid Notice for Surplus Vehicle &#8211; February 19, 2026">…

60,611 bytes626 lines

Real sample of rows we keep up to date

Source: profiles.health.ny.gov · directory hospitals · 219 rows · read 2026-10-12. This is a sample; yours starts from your page.

Sample rows from profiles.health.ny.gov · directory hospitals
Name, address	City	Phone	Source	Checked
A.O. Fox Memorial HospitalOne Norton Avenue	Oneonta, NY 13820	Tel: (607) 432-2000	profiles.health.ny.gov · directory hospitals	2026-10-12
A.O. Fox Memorial Hospital - Tri-Town Campus43 Pearl Street West	Sidney, NY 13838	Tel: (607) 561-2021	profiles.health.ny.gov · directory hospitals	2026-10-12
Adirondack Medical Center-Lake Placid Site203 Old Military Road	Lake Placid, NY 12946	Tel: (518) 523-3311	profiles.health.ny.gov · directory hospitals	2026-10-12
Adirondack Medical Center-Saranac Lake Site2233 State Route 86, P.O. Box 471	Saranac Lake, NY 12983	Tel: (518) 891-4141	profiles.health.ny.gov · directory hospitals	2026-10-12
Albany Medical Center - South Clinical Campus25 Hackett Boulevard	Albany, NY 12208	Tel: (518) 262-1200	profiles.health.ny.gov · directory hospitals	2026-10-12

Narrow screens show the first column, the source and the checked date; the CSV carries every column.

CSV sample rows as JSON 196 source pages tracked · 28 fully verifiedA source is fully verified only when every row contains every field it declares. Pages that stop working are listed as not working.

Rows once, then kept up to date.

Send the URL, or describe the list.
Review a sample before you commit.
Get the same columns again after the page changes, weekly if you want.

Quiet failures become visible.

Fewer rows than the previous run
A column that stopped filling
A page that stopped answering

The usual failure is not an error. It is a plausible result that quietly got smaller.

For developers

Two calls.

Get your key: POST /v1/keys
Make a request: POST /v1/extract

Example

curl -X POST https://www.licium.ai/v1/extract \
  -H "Authorization: Bearer ak_your_key" \
  -d '{"url": "https://jobs.example.com/acme",
       "fields": ["title", "location", "url"]}'

{ "success": true, "data": { "count": 143, "rows": [ ... ] } }

Rows come back as JSON with the source and the checked date on each. A page we do not keep yet returns a request number and a bounty board link, so there is always a next step.

For agents

Same door, through MCP.

Connect once: https://www.licium.ai/mcp
Request, then read: request_source · get_rows

The answer carries the same three outcomes. Or start from one file: curl -s https://www.licium.ai/skill.md. On this page an agent can fill the request through fill_extract_request_form; a person still clicks.