Skip to content

WorkWingman — Company Research (Plain Language)

Before an interview (or even before you finish your cover letter), it helps to know a few things about the company: what do they actually do, what technology do they build with, and has anything newsworthy happened to them lately? WorkWingman gathers this automatically — with the same hard rule as everything else in this app: only real information from free, public, official sources. Nothing made up, ever.

What you get

For each job's company, once it's been synced into your queue, WorkWingman tries to put together a short "interview prep" briefing with up to four things:

  1. What they do — a plain-English summary, usually pulled straight from the company's Wikipedia page.
  2. Tech stack — programming languages and tools the company's engineers actually use in their public work on GitHub, plus anything the job posting itself calls out (like "we use Kubernetes and Snowflake").
  3. Recent news — a few of the most recent news headlines about the company, each one linked back to the original article so you can read the full story yourself.
  4. Talking points — a couple of short, factual sentences built from the three things above, worded so you could drop them straight into a cover letter or bring them up in an interview ("Their public engineering org is heavily Python" or "Recently in the news: ... ").

If a company has nothing available in one of these categories — say, they're a small private company with no GitHub presence — that section just doesn't show up. WorkWingman never invents a description, a technology, or a news story to fill the gap.

Where each piece comes from

  • What they do comes from Wikipedia's own summary API — the same kind of short paragraph you'd see at the top of a Wikipedia article. It's free, official, and doesn't require any account or API key. If a company doesn't have a Wikipedia page (true for most smaller or private companies), this section is simply left blank.

  • Tech stack comes from two places working together:

  • GitHub, if the company has a public GitHub organization — WorkWingman looks at what programming languages show up most across their public code repositories. This only sees what the company chooses to make public; it can't see private/internal codebases, and it can miss companies whose GitHub account doesn't match their company name exactly.
  • The job posting itself — WorkWingman also just reads the job description you're applying to and picks out any technology names it recognizes (things like Python, AWS, Kubernetes, Snowflake). This part never touches the network at all — it's just reading text that's already sitting right there in the job listing. When both agree — say the job description says "Python" and the company's GitHub org is also mostly Python — that's called out as an especially strong signal, since it's real, cross-checked evidence rather than a guess.

  • Recent news comes from GDELT, a free, open news-monitoring project (funded for research, not a commercial product) that indexes news coverage from around the world with no signup required. WorkWingman searched for a second option too — Google's own news search feed — and confirmed it works and is legal to use for this kind of personal lookup, but chose GDELT as the primary source since its open research-project license is a cleaner fit than Google's feed (which is meant for personal reading, not being built into an app). If a future version needs a second news source, that Google option is documented and ready to add.

  • Talking points aren't fetched from anywhere — they're built locally, on your machine, by combining whatever real data came back from the sources above into short, readable sentences. If nothing came back from any source, there simply are no talking points; WorkWingman never papers over a lack of data with a generic, made-up sentence.

What we looked at and decided not to use

A few well-known company-research tools were considered and set aside, all for the same reason: they require a paid plan or a locked-down API key, which doesn't fit WorkWingman's "no accounts, no subscriptions, free to run" design:

  • NewsAPI.org and Bing News Search (paid/keyed news APIs)
  • BuiltWith, Wappalyzer, and StackShare (paid or access-restricted tech-stack lookup tools)
  • Crunchbase (paid company-data API)

None of these have any code in WorkWingman today — they're just written down as "considered, not used, and why" so a future contributor doesn't have to re-research the same question. Wikidata (a more structured, database-like cousin of Wikipedia) was also considered for "what they do," but Wikipedia's own summary already gives a ready-to-read paragraph with less technical overhead, so that's what's actually wired in.

How it feeds your cover letter

The same research that builds your interview-prep briefing also quietly improves your cover letter. When WorkWingman drafts a cover letter for a job, it pulls in up to two of the "talking points" from that company's briefing and weaves them into the letter as a short paragraph — a genuine, factual detail about the company, not a generic filler sentence. If there's no research data for that company yet, the cover letter is drafted exactly as it always was — nothing breaks, nothing looks different, you just don't get the extra personalized detail.

Where you'll see this

There isn't a dedicated "company profile" page yet, so today this shows up as a card on the Documents screen, right alongside your tailored resume and cover letter, whenever a job has research data available. It was built as its own reusable piece, so it can be dropped onto any future screen (like a full job-detail page) without extra work.

Privacy

Every lookup goes out to a public, official API (GitHub, GDELT, Wikipedia) — nothing more than the company's name is sent, which is already information visible right there in the job posting. Nothing about you personally is sent anywhere. This research happens once per company, when a job is first added to your queue — WorkWingman doesn't keep re-checking the same company over and over in the background.