LightningApply
Job applications at lightning speed
I applied to 200+ jobs over 10 months. Every one of them wanted a slightly different resume, so I kept rewriting the same bullet points to match whatever that particular job description emphasized.
Around my 100th application I stopped bothering. Started sending the same generic resume everywhere, because tailoring took me close to an hour and I had zero evidence it was doing anything.
That's the real problem with tailoring. It isn't hard. It's boring enough that you quit.
So I built LightningApply: https://lightningapply.com. My first version didn't even have a UI. It was literally a JSON file of my experiences and skills, a TXT file of the job posting, and a CLI command I'd run to create a resume tailored to it, using Ollama. Took 10 weeks of nights and weekends mostly spent fighting prompts, but the first time it gave me something I'd have actually sent, I just sat there for a second, because I'd gotten an hour back.
Then I kept using it on my own search and kept running into follow-up questions. How well do I actually match this role? What's missing? What do I already have that I'm burying on page two? Each of those turned into a feature, mostly because I wanted it at 11pm on a Tuesday and nothing else would tell me.
I did use it on a real application and landed an interview the next day. No way to know how much of that was the resume. But it was the point where I stopped thinking of this as a script and started thinking of it as a product.
Stack and numbers:
Laravel API, React frontend, Cloudflare, Railway, Postgres, MailerSend
$20/month to run the whole stack
0 users. Launching today
Subscriptions with a free trial to try it out
24 weeks from first commit to launch
What I got wrong: Started on Ollama running on my own machine, where inference was free and I didn't think about it once. Moving to hosted models is what eventually broke the credit model. I tried going for credit-based billing from the beginning but the more AI features I added to the product, the more I realized it wasn't sustainable. Those costs needed a ceiling and the usage wasn't entirely clear to the user.
And that's the thing I'm still not settled on. Moving to subscriptions fixed my cost problem, but it created a different one: job searching is bursty. You apply hard for six weeks and then don't touch it for a year. A monthly subscription is exactly wrong for that shape of usage, and I expect brutal churn the month someone accepts an offer.
If you've sold to bursty users, what worked? Seasonal pricing, a pause option, an annual plan priced as if they'll only use two months of it?
Happy to answer anything about the build.
6 Comments
"Lightning speed" attracts the desperate job seeker who applies to 100 jobs a day with a generic resume. The buyer who pays is the career coach or coding bootcamp who wants to offer application automation as a tool to their 200-student cohort. The individual job seeker has $8/mo willingness to pay and churns after hire. The institution has $500/mo budget and recurring need. Your page speaks to the individual. The bootcamp buyer never finds you because they search "tools for my students," not "apply faster."
That’s actually a great point. Maybe there’s an option to “scale” it for hiring/career coach orgs! Thanks for the feedback on that.
Yeah exactly — and the interesting part is the product itself probably doesn’t need to change much, just the framing + entry point.
Right now it’s positioned as “help me apply faster,” which pulls in low-intent, high-churn users.
But for a bootcamp or coach, the value is completely different — it’s outcomes at scale (placements, success rates, differentiation).
There’s probably a version of this where:
• the individual tool becomes the demo
• but the real product is “a placement acceleration layer” for cohorts
That shift alone can 10x pricing and retention without adding much complexity.
If you’re open to it, I can sketch how I’d reposition this to land those org buyers — feels like a pretty clear upgrade path here.
The bursty usage point feels like the harder problem here.
Do you think people are actually paying for ongoing access, or for a concentrated period where they’re trying to get from applications to interviews?
Agreed. Yeah my whole philosophy was sign up for the search and cancel when you land the next role. Not great for MRR but pretty much the exact reason I wanted to go a credits-based route.
That makes sense. The decision to design around the customer eventually leaving, rather than trying to manufacture retention, is actually the part that caught my attention.
I'd be interested in continuing the conversation by email if you're open to it. What's the best email to reach you on?