/
Our story
Why we built Talent Approved
Talent Approved didn't start as a product. It started because we needed to screen candidates for a hire of our own - and the tools I'd used before no longer fit how we hire. Here's the whole story.
The hire that started it
We hire long-term contractors on Upwork every now and then - usually software developers. This time we were looking for a QA engineer: someone to run manual regression testing, manually test new features, and build automated end-to-end tests with Playwright.
The role also meant living in Jira and owning the gate for new features merging into our main branch - and I wanted whoever we hired to lean on AI tools like Claude Code to move faster, with solid AI-assisted development habits. It's a specific mix of skills, and that mix turned out to matter.
How we were already hiring
Our process was hands-on. We posted an invite-only job on Upwork, then I manually searched for candidates who matched on the things we care about - language, timezone, experience, and rate.
I saved the 30 best profiles to disk and had AI rank them into a single table so I could compare them properly. Then we invited the top 10-15, and the job post told them the next step was a short skill test before any interview. (The interview also helps us gauge English, since we live in standups and daily chat.)
Where the tools we'd used fell short
I planned to use TestGorilla, like I had before. But when I came back to it, the plan I needed had moved to an annual commitment - and we don't hire on a yearly schedule, we hire in bursts. Paying to keep a subscription warm between hires made no sense.
So I turned to Testlify, which I'd also used before. This time I couldn't find an off-the-shelf test that matched the exact combination we were hiring for - manual QA, plus Playwright, plus an AI-assisted workflow. Always close, never quite right.
So I built the test myself - then the whole platform
I thought: how hard can it be? I'll just build the test myself. So I described the role, generated the questions, refined them - and one thing led to another.
That experiment became Talent Approved: describe a role or paste a job description, let Magic Create draft a tailored test, invite candidates with a link, and review results with AI summaries and anti-cheat recordings. No subscription and no per-seat fees - just $5 per completed candidate, and your first 3 are free. It's the tool I wished we'd had - and now it's the one we use for our own hiring.
Who it's for
Built for people who hire like us
Talent Approved is for people who hire occasionally and for specific roles - founders, small teams, agencies, and recruiters who don't want a year-long platform contract for a handful of hires. It's made in Sweden, and it's the same tool we use for our own hiring.
Thanks for reading - and good luck with your next hire.
Jimmie Antonsson
Founder, Talent Approved
Connect on LinkedInFrequently asked questions
Is this a real product or a side project?
It's a real, working product - it started from a real hiring need of ours, and we use it for our own contractor hires.
Why pay-per-candidate instead of a subscription?
Because that's the model I wanted as a buyer. We hire in bursts, not on a yearly schedule, so $5 per completed candidate - with no subscription and credits that never expire - fits occasional hiring far better than an annual plan.
Do I have to be hiring developers to use it?
No. The test is generated from whatever role you describe, so it works for support, sales, virtual assistants, analysts, and other roles just as well as for developers.
Is it actually cheaper than the alternatives?
For occasional or per-role hiring, usually - there's no annual contract or seat fee, and your first 3 candidates are free. For constant high-volume hiring a subscription's bundled credits can win, so it depends on how often you hire.