NewSearch millions of jobs from your AI agent with MCP

How does TheirStack work?

TheirStack works by collecting job postings from several hundred thousand websites, including company career sites, applicant tracking systems and job boards, and reading each one for the technologies it names. Mentions are rolled up per company into a technology profile with a confidence score, and the jobs, companies and technologies are served through a web app, an API, webhooks, an MCP server and bulk datasets.

Collection uses publicly available data. TheirStack's docs say it scrapes some websites directly and reaches others through job board aggregations, crawls continuously rather than on a nightly schedule, and lists Indeed and LinkedIn as its two largest sources by share. Each job is stored once, the first time it is found, with the date it was posted, the date it was discovered and the date it closed, and a repost of the same title by the same company within thirty days is ignored. Detection then looks for a technology in three places: the job title, the description, and the job URL, which reveals the applicant tracking system a company uses. The confidence score combines how many postings mention the technology, how that compares with rival tools in the same category, where the mention appears and how recent it is.

JobsPipe is built on the same raw material and exposes it at a lower level. POST /v1/jobs/search returns normalized postings from named sources, filtered with job_title_or, company_name_or, skills_or, description_or and posted_at_gte, and each posting carries a technology_slugs field parsed from its text. Grouping the results of a skills_or search by company gives a job-posting technographic list you assemble yourself. POST /v1/stack/scan uses a different mechanism, fingerprinting what a company's domain serves, so the two methods can be cross-checked.

To see the mechanism rather than read about it, open a technology card in the TheirStack app and follow the link to the postings that produced it. For the developer view, send a skills_or query to the JobsPipe sandbox search endpoint, which needs no key, and read the postings that come back.

The full treatment is on Job data vendors and tools.

Related questions

Where does TheirStack get its data?

From public job postings. Its docs describe several hundred thousand source websites, made up of company career pages, applicant tracking systems and job boards, with Indeed and LinkedIn the largest by share, followed by sources such as Workday, Naukri and SmartRecruiters. Some sites are scraped directly and others are reached through job board aggregations. Technographic data is derived from those same postings, not from a separate crawl of company websites.

Does TheirStack scan company websites like Wappalyzer or BuiltWith?

Not according to its methodology page, which says its technology usage data is sourced from job postings. That is the opposite trade-off to a website scanner: it can surface internal tools such as a CRM, an ERP or a data warehouse that never touch a public page, but it misses companies that do not hire for a tool. A domain scan such as JobsPipe's POST /v1/stack/scan covers the website side.

How does TheirStack handle duplicate job postings?

Its docs say it deduplicates aggressively. When the same title from the same company reappears within thirty days, the first posting is kept and the repost is ignored, which also prevents the same job being charged twice in credits. A job found on several sites is stored once, and the source field records where it was seen first, so the docs advise against reading that field as where a job was published.

Does TheirStack have an API and webhooks?

Yes. TheirStack documents a Job Search endpoint, a Company Search endpoint and a Technographics endpoint, plus webhooks that fire on events such as a new job, a closed job or a new company, and an MCP server for AI agents. API credits are consumed per record returned or per webhook event dispatched, and the cost is the same either way, so repeated requests for the same data are billed each time.

More answers

Query live postings yourself, 1,000 jobs a month free.

Get a free API key