Gregory Bolshakov

Every company on Welcome to the Jungle: sector, size, offices, tech stack, open advert count and the company's own website. The site's public endpoint returns the same 100 companies whatever page you ask for. The index holds 17,189 records, which are 12,423 real companies, each returned once.

Every French professional equality index published on Welcome to the Jungle, with all five indicators split out. The indicator fields are POINTS SCORED and not the size of a gap, so a pay gap field of 40 is a full mark. No filter exists for the score, so profiles are read to find it.

All 88,651 adverts on Welcome to the Jungle: salary, remote policy, benefits, profession, and the employer's own application link. A quarter of salaried adverts are quoted per month, so the raw number understates them twelvefold. This one publishes the period beside it.

Every office of every company on Welcome to the Jungle, one row per site with street address, postcode, city, country and map coordinates. The street and the coordinates are only on the per-company profile, and the search index ships coordinates in a separate array that is not always aligned.

Every company in the Y Combinator directory: batch, industry, status, team size, tags and locations. YC search stops at 1,000 results and reports the stop as an empty page, so a plain scraper returns 1,000 of 6,203 and calls it done. This one slices by batch and gets the rest.

The fields the YC directory search does not have: founding year, city and country, LinkedIn, Twitter, Crunchbase and GitHub, the YC group partner, press items, launch count and the real number of open jobs. Filter by batch, industry, region or tag, or name the companies.

One row per founder of a Y Combinator company: name, title, bio, LinkedIn and Twitter, with the company beside them. The directory search YC publishes holds no people at all, so this reads the company pages. Filter by batch, industry, region or tag, or name the companies yourself.

Every open job at Y Combinator companies, read company by company, with pay parsed into currency and period. The YC jobs pages ignore their own filters and cap at 50 with no total, so a landing page is a sample. Filter by batch, industry, region or role, or name the companies.

Every Launch YC post with its whole body, upvotes, date and the company that shipped it. YC search hands back 1,000 of 3,271 and reports the stop as an empty page, so this slices by batch and gets the rest. Image links in the body are rewritten to full URLs.

Open procurement notices from TED, the EU's official journal. Groups every notice about the same procurement into one row and takes the deadline from the newest one, so a corrected tender is not three rows with three closing dates. No key and no login.

Searches US federal disaster declarations from OpenFEMA. FEMA returns one row per designated county, not per disaster, so a row count is not a disaster count. This Actor gives you either grain and always reports both numbers. No key and no login.

Search every public repository on GitHub by topic, language or stars, or watch for new ones on a schedule. Gets past the 1,000 result ceiling by splitting the creation date range, and every row says how many matches GitHub has so you know what you did not get. No key and no login.

Searches US federal funding opportunities on Grants.gov and returns the agency contact, the award amounts and the real closing date. Grants.gov keeps two records per opportunity and the older one often has the higher version, so every row says which one its deadline came from.

Search every Hacker News story and comment, or watch for new mentions of your company on a schedule. Gets past the 1,000 result ceiling by splitting the date range. No key and no login.

Searches the NPPES NPI Registry for US doctors, dentists, clinics and other healthcare providers, with practice address, phone and specialty. NPPES caps any one search at 1,200 records and never says how many matched, so every row here carries whether the answer is complete. No key and no login.

Watches any number of RSS, Atom or RDF feeds and returns their items, flagging the ones that were not there last run. Tells a feed that refused the request apart from a feed with nothing new, which is the difference between a monitor and a guess. No key and no login.

Screen names against the OFAC SDN, OFAC consolidated, EU and UN sanctions lists. Every row says which name matched, whether the publisher calls that alias weak, and which lists were actually searched. No key and no login.

Reads Steam game reviews with playtime, votes and author history. Steam has two defaults that each narrow what you get, language and purchase type, and its own total shrinks to match. Apex Legends reports 1,044 reviews that way and has 1,066,930. This asks for all of both. No key, no login.

Searches US federal contract, grant and loan awards on USAspending.gov. Its paging flag turns false at 10,000 records while tens of thousands more are still there, so this Actor stops on a short page instead and uses the cursor to read past 50,000. No key and no login.

Checks EU VAT numbers against the Commission's own VIES service and returns the registered trader name and address. Tells an invalid number apart from a member state that did not answer, which the raw service does not. No key and no login.

Hourly weather for any place, by name or coordinates, up to 16 days ahead and 92 days back. Every row carries both the local time and a real UTC instant, so readings from different locations are actually comparable. No key and no login.

Get every reply to any X post. Official X API, you pay only for replies you receive, and long jobs do not time out.