Most auto-apply tools call it sent when they click submit. We wait for the employer to say so.
AI Applyd is live on Peerlist's Launchpad this week, so here is the part that actually took the longest to build.
Every tool in this category reports success off its own click. The user sees "submitted", the employer never received anything, and nobody finds out for weeks. We decided an application does not count until the employer's own system confirms it arrived, and then we had to go and build that.
What the product does now:
- You swipe through a deck of jobs matched to what you actually want, not a keyword dump.
- Swipe right and it rewrites your resume for that specific posting and writes the cover letter in your voice.
- It opens the employer's real hiring form on Greenhouse, Lever, Ashby, Workday, iCIMS, Personio, Teamtailor, SmartRecruiters, Recruitee, Breezy, Workable or Rippling, and submits.
- It gets through account walls and captchas by itself. You are never asked for a code or a login.
- The application only turns to sent when the employer's system confirms receipt. When it cannot confirm, it says so and tells you why.
The hard part was never the form filling. It was deciding what counts as proof, and then refusing every weaker signal: our own click, an HTTP 200, an internal status, a page that merely looked like a thank-you. Only a linked employer receipt moves a row to sent.
It also runs inside Claude, Cursor and ChatGPT over a remote MCP server, which is how a lot of people find it now. Same tools, no install, no API key.
Try it on one job you actually want and watch the application get written in front of you.
I build AI Applyd - https://aiapplyd.com?ref=indiehackers - auto-apply that proves your application arrived.
What would make you trust that an application actually landed? That question keeps changing this product, so I would rather hear it than guess.
Adding the link, because a few people asked where to actually look.
We are live on Peerlist Launchpad all week: https://peerlist.io/firstexhotic/project/ai-applyd-mcp
If any of this is your kind of problem, come and say what you would change. A comment there reaches me directly and it is worth more to us today than anything else.
This is a much better definition of success than trusting the click event. How are you handling the cases where the employer confirms receipt but the application gets filtered before a human sees it?
Honest answer: a receipt proves delivery, not that a person read it. Those are two different problems, so we treat them separately.
Before the send, where we can actually move the odds: it only puts you forward for roles you genuinely fit, and it rewrites the resume in that posting's own language, so the filter reads the fit instead of missing it. That stage is where most applications quietly die.
After the send, where we can observe: replies come back onto the same application as the receipt. A rejection, an interview invite, or plain silence each show up there instead of getting lost in your inbox. Silence past a sensible window gets a follow-up draft, not another blind apply.
What we won't do is pretend a receipt means a recruiter opened it. If you've found a reliable signal for "a human actually looked at this", I'd love to hear it.