Senior-level analysis, every time
The AI is prompted to act as a senior HR consultant with 15+ years of experience. It reasons about intent behind requirements — not just keyword matches — so coachable gaps don't disqualify strong candidates.
Powered by LLaMA 3.3 70B on Groq · Free to use
AI resume screening
that thinks like a senior recruiter
Upload resumes, paste a job description, and get a structured candidate analysis in seconds — match score, hire / interview / reject verdict, red flags, and tailored interview questions, streamed live.
No signup GDPR / RODO compliant No data stored EN + PL
hr-ai-assistant.app
Candidate
Anna K.
92%
Hire
Candidate
Michał P.
74%
Interview
Candidate
Tomasz W.
58%
Pass
< 200 ms
First token latency
~500 tok/s
Groq inference speed
20+
Resumes per batch
2 langs
EN + PL native
Features
Everything a senior recruiter does, in seconds
Match scoring, deep analysis, follow-up chat, multi-candidate batches, and rigorous GDPR compliance — built into a single, fast, glass-clean interface.
Real-time streaming with Groq
Powered by LLaMA 3.3 70B on Groq inference (~500 tok/s). First token in under 200 ms, full analysis in seconds — not the 30-second wait you'd expect from a deep review.
Threshold-enforced recommendations
Every candidate gets a precise match percentage and a binding decision: Hire (≥80%), Invite to Interview (65–79%), or Do Not Hire (<65%). No ambiguous fence-sitting.
Multi-candidate batch screening
Drop in 5, 10, or 20 resumes. Each is parsed in parallel, scored against the same job description, and surfaced with sortable summary cards — pick your shortlist in minutes.
Per-candidate follow-up chat
After analysis, every candidate gets a dedicated chat. Ask why a score was given, request tailored interview questions, or probe a specific gap — the AI remembers full resume context.
Bilingual interface, any-language CVs
Full English / Polish UI toggle. The AI replies in your selected language by default but adapts on the fly if you write in another — perfect for cross-border hiring.
GDPR & RODO compliant by design
Resumes are parsed in memory, sent to the LLM, and discarded. Nothing is stored, logged, or used for training. Explicit consent flow and full Polish / English privacy policy.
PDF & DOCX, parsed server-side
Drag-and-drop PDF or DOCX. Files are extracted to plain text on the server before reaching the model — your candidate's raw file never touches a third party.
Structured, actionable reports
Match score, strengths, red flags, trainability prognosis, cultural fit, 3-month expectations, and a 2–3 sentence final verdict with the next concrete action. Skim or deep-dive.
How it works
Four steps. Under a minute.
No setup, no integrations, no learning curve. The whole tool is a single page — the only thing between you and a shortlist is the upload button.
01
Paste the job description
Drop in the role spec — anything from a one-paragraph brief to a full requisition.
02
Upload resumes
PDF or DOCX, single or batch. Add optional notes per candidate (referrals, interview feedback, anything relevant).
03
Read streaming analysis
Match percentages, hiring recommendations, strengths, red flags, and tailored interview questions appear in real-time.
04
Chat with each candidate's report
Drill into any analysis with a per-candidate chat panel. The AI keeps full resume context and answers follow-ups instantly.
Scoring philosophy
Optimistic-by-design.
No keyword theatre.
Most ATS tools penalize candidates for missing exact keywords — a senior data scientist who "doesn't have scikit-learn" because their CV says "production LLM pipelines" gets dropped. We built the opposite.
The model is instructed to treat false negatives — overlooking a promising candidate — as more costly than false positives. It reasons about intent behind requirements, classifies skill gaps as critical / coachable / optional, and rewards strong learning trajectory.
Tiered gap assessment
Critical
Core hard skill with no adjacent proof of competence
−20 to −30%
Coachable
Missing a specific tool, but strong in the domain
−5 to −10%
Optional
Nice-to-have requirement not met
−2 to −5%
Hiring thresholds
≥ 80%
Hire
65–79%
Interview
< 65%
Pass
Built for
Anyone who reads resumes for a living
In-house recruiters
Screen 50 inbound resumes for a single role in the time it used to take to read three. Surface your shortlist, then go deep with the per-candidate chat.
Hiring managers
Get a senior HR opinion on a candidate before your first call. Pull tailored technical questions for exact skill gaps, ready to use in the interview.
Recruitment agencies
Triage candidate pipelines across multiple client roles in parallel. Ship structured shortlists to clients with reasoning they can audit.
Startup founders
Hire without a recruiter. Get a structured opinion on every applicant, in plain language, without paying $300/seat for an enterprise ATS.
Resumes parsed in memory. Nothing stored.
CV files are converted to text on our server, sent to the LLM for analysis, and immediately discarded. No database writes, no logs of resume content, no training on your data. GDPR & RODO compliant by design — with a published privacy policy and explicit consent flow in both English and Polish.
GDPR readyRODO readyNo data retentionServer-side parsingRate-limitedStrict CSP headers
FAQ
Frequently asked questions
How does the AI resume screening work?
+
Paste a job description and upload one or more resumes (PDF or DOCX). The tool parses each file server-side, sends the text plus the job spec to a large language model, and streams back a structured analysis: match percentage, hiring recommendation, strengths, red flags, prognosis, and interview questions. Results begin appearing in under a second.
Is HR AI Assistant free to use?
+
Yes. The core analysis tool is free to use. There is a sliding-window rate limit (10 requests per minute per IP) to protect against abuse, but no signup, paywall, or trial restriction.
Is it GDPR / RODO compliant?
+
Yes. Resume content is processed server-side only for the duration of the analysis and is not stored, logged, or used to train models. The interface is fully translated to Polish (RODO) and English (GDPR), with explicit consent and a published privacy policy.
What file formats are supported?
+
PDF and DOCX. Files are parsed server-side using pdf-parse and mammoth before any AI processing — the LLM only ever sees plain text, never the raw file.
Can I analyze multiple candidates at once?
+
Yes. Batch upload as many resumes as you need; each candidate is assigned a numbered slot and analyzed in parallel against the same job description. After analysis, each candidate gets a private chat panel for follow-up questions with full context.
What languages are supported?
+
The interface is bilingual — English and Polish — and the AI adapts to whichever language you write in, including resume content in any language. Switch the UI language with one click.
Will this replace our ATS?
+
It complements your ATS rather than replacing it. Use HR AI Assistant for the screening and shortlisting layer — the part where most teams spend the most time — and keep your ATS for pipeline, scheduling, and offers.
How accurate is the match score?
+
The model is tuned with an optimistic-by-design philosophy: it treats false negatives (overlooking a promising candidate) as more costly than false positives. Skill gaps are tiered — critical, coachable, optional — with proportional score impact, so a strong learning trajectory isn't penalized for missing a specific tool.
Stop reading resumes.
Start interviewing.
The whole tool is one page. No signup, no credit card, no setup. Open it, upload, and you'll have a shortlist before your coffee cools.
Loved by 100+ early users