Launched on Product Hunt yesterday after 18 months. 4 upvotes in. Asking for honest support.
Quiet launch yesterday. Self-hunted, sole maker, eighteen months in. Four upvotes and one comment as of right now. Posting on IH because this community is the one place I can actually say that out loud.
AI Applyd is a job-application agent. Paste a job URL, it scores your resume against the JD's ATS keyword extraction, rewrites the resume to fix the gaps, then submits the application end to end on Greenhouse, Lever, Workday, Ashby, iCIMS, and LinkedIn Easy Apply.
Free tier: 20,000 tokens, 10 ATS scores per month, 5 resume tailors per month, 1 fully covered auto-apply per month. No credit card.
Things I learned shipping this:
-
The agent-browser part is harder than the AI part. Workday alone took 4 rewrites and the failure rate is still meaningful on certain custom flows. Free tier eats those costs. Return the unused token reservation and tell them to apply manually rather than fake a success.
-
Free-tier infra is the only way solo. Edge compute, edge database, edge storage, free model tiers for system paths. Paid models only on user-triggered apply. Took 2 surprise bills to get this discipline.
-
Honesty in the marketing copy converts better than confidence. The PH page openly says what is still rough. Early signups don't churn as hard.
PH link: https://www.producthunt.com/posts/ai-applyd-2
If you have 30 seconds and a PH account, an honest upvote helps. Not asking for anything more than that. Happy to answer anything about the stack, the failure modes, or the pricing math.
Hey Ava, 18 months of building and now you're out there -- that takes real dedication. Hope the upvotes start rolling in! Wanted to share something that might help keep the momentum going. There's a tool called Markey (markey.app) that keeps your product visible on social consistently after launch so you don't go quiet after the PH spike. Might be exactly what you need right now. Good luck, rooting for you!
appreciate the kind words, genuinely. i'm wary of 'stay visible after launch' tools by reflex, mostly because consistency is the one thing i've never managed to outsource to a product, it always comes back to actually showing up. but i'll take a look. thanks for rooting.
Hey Ava, stumbled upon this post so figured i'd give my feedback:
As someone with 10+YoE in engineering, I personally don't think I'd ever use this, especially now in the AI era.
By no means am I saying this product is bad (looks great), but there are many examples of recruiters / companies moving away from inbound applications given the huge influx of automated applications. This worries me a bit and makes me think this product will be an uphill battle for you. There are already a ton of bots that can auto-apply for you. I also think it might be challenging to to get real conversions from AI-automated applications (maybe a small % will get lucky?).
Most of my job sourcing comes direct from recruiters through my network, and I feel pretty strongly that the future of job hunting will be through trust networks, not automated bot applications. If I do send an application (very rarely), I make sure to personalize the intro as people are getting really good at sniffing out AI.
If I were you (take this with a grain of salt, please), I would rethink this product a bit. If your product was building a way to weed through the overwhelming amount of ai slop in the world right now, I'd be more inclined to use it.
Just my $0.02.
this is the most useful comment i got on that post, so thanks for not softballing it. you're right that the inbound flood is real and companies are actively defending against it. the way i see it, that flood is exactly why 'did it actually submit, and did a human ever read it' matters more than raw volume. a tool that fires 200 applications into a void is part of the problem you're describing, not a fix for it. where i'd push back gently is treating trust networks and applications as either/or. most people early in a search don't have the network yet, and the cold-apply lane is where they're most invisible and most lied to by tools that report success they never confirmed. weeding the slop and surviving the slop are the same problem from two sides. i won't pretend cold-app conversion is great, it isn't, and i'm not going to hand you a number i can't stand behind. appreciate the actual feedback, it's worth more than two cents.