Resume optimization the way a senior recruiter would do it.
HiredNa started in a Manila kitchen during the 2024 hiring downturn. Three engineers, all fielding the same complaint: 'I'm getting ghosted, and I think my resume is fine.' It wasn't fine. It was generic — written for 'Software Engineer' instead of 'Backend Engineer at a payments company.'
We built a simulated recruiter to give honest feedback at scale, a niche specialist that grades you against the corpus of jobs you actually want, and a guard that refuses to invent metrics. The combination is HiredNa.
Today the team is small and distributed across Manila and New York. We ship weekly, publish a public changelog, and answer founder email at hello@hiredna.app. If you want to talk about how we build, why we bet on specialist agents over a single generalist model, or what good hiring practice looks like in 2026 — reach out.
Median pipeline runtime
Niche specialisations
Fabricated metrics tolerated
What we believe
Specialists, not a generalist
Seven narrow agents — parser, analyzer, niche specialist, recruiter, optimizer, guard, renderer — beat one giant prompt. Each has a single job and a single failure mode you can read.
Source-of-truth resumes
Your facts are immutable inputs. The hallucination guard rejects rewrites that invent employers, dates, titles, or metrics. The pipeline halts. You don't ship lies.
PH-first pricing, global rendering
Built in Manila, priced for Manila first. The same agent pipeline serves PH ₱ and US $ tiers — region detection picks the right currency at checkout.
Reviewable diffs, never autopilot
Every bullet rewrite is a hunk you accept or reject with j/k navigation. The model proposes; you decide.
The stack, on the record
Next.js 16 App Router on Vercel. Convex for durable workflows, persistence, and the real-time UI. Clerk for auth. Stripe for billing. Vercel AI Gateway routes between Gemini, Claude, and GPT — agents fall back automatically when a provider degrades.
We use @react-pdf/renderer for ATS-safe PDF output, @convex-dev/agent for typed LLM threads, and a per-skill RAG corpus for the niche specialist. None of it is secret — we'd rather you copy the architecture than pay for a black box.