Tech Stack Adoption Monitor from Job Posts

Find out which companies just started asking for your product's category — before their website shows any sign of it.

Every tech-stack detector on the market scans a company's website. That only reveals the front end: CMS, analytics, pixels, chat widgets.

It cannot see the backend. Nobody puts Snowflake, Kubernetes, Kafka, Terraform or Rust in their HTML. Engineers declare those in one place only: the job posts they write to hire for them.

Signal	What it means for you
🟢 Company adopted	They just started requiring the tool. That is a buying signal
🔴 Company dropped	They stopped asking. They migrated to something else
📈 Market demand up	The technology is gaining ground across companies
📉 Market demand down	It is losing ground

If nothing changed, you get nothing and you pay nothing per result.

What a run looks like

6,655 job posts across 60 companies → 1,224 company/technology pairs → 4 signals:

[tech_adopted] Stripe started asking for Snowflake (2 job posts)
[tech_dropped] Brex stopped asking for Datadog
[demand_up   ] Kubernetes up 105% across the market (18 → 37 companies)
[demand_down ] Terraform down 50% (68 → 34 companies)

Every adoption signal ships with a sample job post and its URL, so you can open it and see the exact wording before you reach out.

No company list needed

Ships with a validated index of 872 companies across Greenhouse, Lever, Ashby and Workable. Point it at your own list instead if you prefer:

{
  "companyTokens": ["greenhouse:stripe", "ashby:ramp"],
  "technologies": ["Snowflake", "Databricks", "dbt"],
  "monitorKey": "data-stack-watch"
}

Watch only your category

70+ technologies across 10 categories: languages, frontend, cloud, infrastructure, data, databases, observability, AI/ML, business SaaS and security.

Sell a data warehouse? Watch data. Sell observability? Watch observability. Far less noise than watching everything.

Built for schedules

Run it weekly. Adoption moves in weeks, not hours — a daily schedule mostly returns nothing.

• Dev-tool and data vendors — account targeting by real stack, not guesses
• Sales teams — reach out the week a prospect starts hiring for your category
• Investors — track technology adoption curves across a portfolio
• Analysts — build a demand time series per technology

First run is free

The first run records the baseline and charges nothing — there is nothing to compare against yet. From the second run onward you get adoption changes only.

Just need the current stack right now?

Set oneShot: true and it returns which technologies each company asks for today, instead of what changed.

{ "oneShot": true, "companyTokens": ["greenhouse:stripe"], "categories": ["data"] }

This is the mode to use for a one-off lookup or from an AI agent. As a pure monitor the first run returns nothing, because there is no history to compare against yet. It never touches your saved monitoring history.

Honest limits

• This measures what companies say they want to hire for, not what they run in production. A job post asking for Kubernetes is strong evidence, not proof.
• Ambiguous names are handled carefully, and some are deliberately strict. Segment only matches "Twilio Segment" or "Segment CDP", because bare "Segment" matched job titles like "Startup Segment" in testing. Rust, Swift and Go require capitalisation or an unambiguous alias like Golang. This trades a little recall for far fewer false positives.
• Keep maxCompanies and your technology list stable between runs. Changing them changes what is being watched and shows up as a burst of false adoptions.
• Market demand signals need at least 3 companies and a 20% shift, to avoid noise from tiny samples.
• Greenhouse job posts carry full descriptions; coverage is best there.

Where the data comes from

The public, unauthenticated JSON endpoints that Greenhouse, Lever, Ashby and Workable publish so companies can embed job boards on their own sites. No login walls, no headless browser. Only job content is read — no personal data about recruiters or applicants.