Help — How to read your ATS report

Score interpretation

Your total score is computed from 5 sub-scores that together add up to 100. Sub-scores are graded on real ATS signals — not vibes.

85+

Excellent

ATS-ready and recruiter-ready. Polish low-severity items only.

70–84

Strong

Will pass most ATS. A few targeted fixes will push you to excellent.

55–69

Decent

Some ATS will struggle. Fix the high-severity items in this report.

40–54

Needs work

Several ATS-blocking issues. Prioritise high-severity fails.

< 40

Major rework

Likely auto-rejected by most ATS. Consider a rewrite.

The 5 sub-scores

Structure & density

/ 30

Contact info present, conventional sections (Summary / Experience / Education / Skills), bullets used, dates per role, reasonable line count.

Industry keyword coverage

/ 20

% of role-specific canonical keywords (from our industry library) found in your resume — variations counted (e.g. "JS" matches "JavaScript").

Readability

/ 20

Action verb leads, quantified outcomes, healthy bullet length, no passive phrasing or first-person pronouns.

Length

/ 15

Word count between 350–1000 (1–2 pages). Recent role has the most detail. Experience longer than Education.

Parseability

/ 15

Whether the file extracts cleanly: not a scanned PDF, no decorative glyphs, dates parseable, no multi-column layouts. This is what determines whether the ATS even gets to read your text.

Anatomy of the report

Top-down, every block on the report explained.

1. Hero card — total score + sub-scores + detected role

Big number on the left is your total ATS score (0–100). The five bars on the right are the sub-scores feeding into it. The pill at the top tells you which band you're in (Excellent / Strong / Decent / Needs work / Major rework).

Detected role + confidence tells you which industry library was used to score keyword coverage. Below ~55% confidence, we fall back to a generic library — your role-specific keywords won't be matched optimally.

2. AI-driven analysis (collapsible blocks)

Generated by our AI engine on top of the deterministic checks. Each block is independently collapsible:

Grammar & spelling — line-level fixes with the original and the corrected version side-by-side.
Rephrased bullets — for every weak bullet, 3 stronger alternatives with a one-line rationale.
Improved Summary / Skills — drop-in replacements you can edit and paste into your resume.
Expert insights — qualitative observations the rules can't catch (e.g. "Most-recent role has 8 bullets but only 1 is quantified").

If you see "AI features are disabled" — your account / deployment hasn't configured an AI provider. The rest of the report is unaffected.

3. All 85 ATS criteria (grouped, collapsible)

Each named check evaluated individually. Grouped by category:

Contact (10) — email, phone, location, LinkedIn, portfolio.
Section structure (12) — Summary, Experience, Education, Skills present and standard-named.
Roles & experience (10) — title / company / dates per role, tense, chronological order.
Bullets (15) — action verbs, metrics, length, no passive phrasing or first-person.
Keywords (10) — industry detection, coverage, no stuffing, hard-skill density.
Readability (8) — sentence length, walls of text, common typos, tense consistency.
Length & density (6) — word count band, estimated pages, section balance.
Parseability & format (10) — text extracted, printable ratio, no columns/tables.
Privacy (4) — no DOB / marital status / sensitive IDs.

Each row shows: status badge, severity, points awarded, evidence (the actual snippet from your resume), and a tip.

4. Section-by-section feedback

One verdict card per resume section (Summary, Experience, Skills, Education, Projects, Certifications). Each card has a verdict — strong / ok / weak / missing — plus a tip.

5. Keyword coverage — found vs missing

Two lists side by side:

Found — canonical keywords from your industry library that we matched in your resume (variations counted).
Missing — the full ranked list (no top-N cap) of role-relevant keywords your resume doesn't yet mention.

Don't blindly add every missing keyword — only add ones you actually have experience with. ATS systems penalise obvious keyword stuffing.

6. Every weak bullet flagged

Every bullet from your Experience and Projects sections that triggered at least one rule, quoted verbatim. The chips above each quote tell you why it was flagged: passive phrasing, no metric, weak verb, too long, too short, first person, etc.

If the AI engine is enabled, the Rephrased bullets block above gives you 3 specific rewrites per flagged bullet.

ATS terminology

The vocabulary the report uses, grouped by topic.

ATS basics

ATS: Applicant Tracking System. Software that screens resumes before a human ever sees them. Examples: Workday, Greenhouse, Lever, Taleo, Naukri, iCIMS.

Parseability: How cleanly the ATS can extract text from your file. Image-only PDFs, multi-column layouts, and tables for layout all hurt parseability.

Recruiter scan: The 6-to-10 second skim a recruiter does before deciding to read a resume properly. Big determinants: action verbs, numbers, recent role at the top.

Confidence score: How sure our engine is about the role it detected. Below ~55% we fall back to the generic keyword library, which means lower keyword-coverage scores.

Resume structure

Section: A labelled block of content (Summary, Experience, Education, Skills, Projects, Certifications). ATS systems look for standard names — avoid clever rebrands like "Where I've Worked".

Summary / Profile: A 2–4 line professional pitch at the top of the resume. Optional but recommended; doubles as the recruiter scan's anchor.

Reverse chronological: Listing roles newest-first. The format ATS engines and recruiters expect.

Role header: The block above each role's bullets — should always have job title, company, and date range. Location is informational.

Standard headers: The exact section names ATS engines recognise: Summary / Experience / Education / Skills / Projects / Certifications. Custom names hurt parsing.

Bullet point: A concise statement (8–25 words) describing one accomplishment in a role. Aim for 3–5 bullets per recent role.

Bullets & impact

Action verb: A strong verb that opens a bullet and conveys ownership: led, shipped, built, designed, owned, drove, scaled. Recruiters skim for these.

Power verb: High-impact verbs (orchestrated, spearheaded, accelerated, pioneered). Use sparingly — too many sounds inflated.

Weak verb: Soft openers like worked on, helped with, assisted in, participated in. They hide what you actually did.

Passive phrasing: "Responsible for", "duties included", "tasked with". Replace with action verbs that show what you owned.

Quantified result / metric: A number, %, $, or scale signal that proves impact. "Reduced p99 latency by 38%" beats "Improved performance".

Scale signal: A scope number — team size, user count, dollar volume — that signals seniority. "Led a team of 6", "4M users", "$2M ARR".

Outcome verbs: Verbs that describe results, not tasks: reduced, increased, improved, grew, cut, saved, generated, accelerated.

Verb diversity: Using a wide variety of action verbs across bullets. Repeating the same verb >3 times signals lazy editing.

Keywords

Industry library: A curated list of canonical keywords + variations for a specific role. We ship 10 industries (Software Engineering, Data Science, Product Management, Marketing, Sales, Design, Finance, HR, Operations, Customer Success) plus a generic fallback.

Canonical keyword: The "official" name we display, even when your resume uses a different variation. If you wrote React.js, the canonical is React.

Variation: An alternate spelling or abbreviation. JS, JavaScript, ECMAScript, ES6 are all variations of the canonical JavaScript.

Coverage %: The fraction of your industry library's canonicals your resume mentions. 40%+ is strong; below 20% the report flags it as high-priority.

Hard skills: Concrete tools, languages, frameworks (Python, Salesforce, Excel, AutoCAD). ATS-friendly.

Soft skills: Behavioural traits (communication, leadership, problem-solving). Use sparingly — let your bullets demonstrate them.

Keyword stuffing: Cramming the same keyword many times in hopes of ranking higher. Modern ATS penalise it; humans see right through it.

Acronym expansion: Spelling out an acronym on first use, e.g. "SEO (Search Engine Optimization)". Helps both ATS and human readers.

Common business metrics

KPI / OKR: Key Performance Indicator / Objectives & Key Results — common goal-setting and measurement frameworks. Mentioning either signals goal orientation.

ARR / MRR: Annual / Monthly Recurring Revenue. SaaS revenue metrics — common in product, sales, customer success bullets.

CAC / LTV: Customer Acquisition Cost / Lifetime Value. Marketing & sales economics metrics.

NPS / CSAT: Net Promoter Score / Customer Satisfaction. Customer-experience metrics.

p95 / p99 latency: The latency at the 95th / 99th percentile of requests. Engineering-performance metrics.

CI/CD: Continuous Integration / Continuous Deployment. Software-engineering practice — automated builds, tests, deploys on every commit.

Status & severity badges

Two pills appear on every criterion: a status (did you pass?) and a severity (how much it matters).

Status

pass Met fully. No action needed. Full points awarded.
partial Almost there. Some criteria met, room for improvement. Reduced points.
fail Not met. Action needed. Zero points.

Severity

high High impact. Likely to block the resume in many ATS. Fix first.
med Moderate. Hurts perception or scores noticeably. Fix after the high-severity items.
low Polish. Quality-of-life improvement. Fix last.

How to use the report

Top-down: read it the way a recruiter would scan a resume.

1
Read the total + sub-scores first.
The total tells you the band. The sub-scores tell you which dimension to focus on.
2
Scan the AI rephrase block.
Each weak bullet gets 3 alternatives. Pick the strongest, paste into your resume, edit to fit.
3
Filter the 85 criteria to "high" + "fail".
Use the collapsible groups. Fix every high-severity fail before touching anything else.
4
Pick 5–10 missing keywords to add.
From the "Missing — full ranked list", pick canonicals you actually have experience with. Add them to your Skills section or weave into a relevant bullet. Don't stuff.
5
Re-export as a text-based PDF.
In Word: File → Save As → PDF. Avoid scanning a printout. Avoid columns and tables.
6
Re-score for free.
The free score is unlimited. Most users see a 15–25 point lift in the first revision cycle.

A concrete action plan

If you're stuck on where to start, do these in this exact order.

Fix every high severity fail in Contact, Sections, and Roles. These block parsing.
Fix bullets that fail on action verbs or metrics. Replace passive phrasing; add a number to every Experience bullet that doesn't have one.
Add 5–10 missing keywords to your Skills section — only ones you actually use.
Apply AI rephrase suggestions to your weakest bullets. Edit them to your voice.
Replace your Summary with the AI-improved version (or a hybrid of it + yours).
Polish the med severity items. Tense consistency, bullet length, terminating punctuation.
Final formatting check. Single column, standard fonts, no tables for layout, no headers/footers.
Re-export as PDF (text-based). Re-upload. Re-score.

Common ATS mistakes

The silent killers — most resumes have at least three of these.

Scanned PDF instead of exported PDF.

A scan is an image. ATS engines can't read images without OCR — and most don't OCR. Always export from your editor.

Multi-column layouts.

Many ATS read top-to-bottom only. Two-column resumes get jumbled.

Tables for layout.

Cells get fragmented during text extraction. Use plain paragraphs and bullets.

Contact info in headers / footers.

Some ATS strip headers and footers entirely. Put email and phone in the body, near your name.

Decorative glyphs (★, ◆, ✦) and fancy fonts.

They confuse parsers and rarely render in the recruiter's preview pane.

Photos, DOB, marital status.

In many regions these create bias risk and are explicitly discouraged. Always omit.

Keyword stuffing.

Repeating the same keyword 6+ times to "rank higher" is detected and penalised by modern ATS.

First-person pronouns ("I led", "my team").

Resume bullets are written without pronouns. Lead with verbs.

Inconsistent date formats.

"Jan 2022 – Present", "2018 - 2020", and "01/2021" mixed across roles. Pick one and stick to it.

Custom section names.

"Where I've Worked" instead of "Experience". ATS look for the standard names — don't be clever.

AI sections — what they are (and aren't)

The AI block sits on top of the deterministic rule engine, not in place of it. Here's what each block does.

Grammar & spelling

Conservative line-level fixes: typos, subject-verb agreement, tense errors, punctuation. Each issue shows the original line and the corrected line so you can decide whether to apply.

Rephrased bullets

For every weak bullet flagged by the rule engine, the AI returns 3 stronger phrasings with a one-line rationale per option. The AI never invents new metrics — where a number would help, you'll see "X%" or "<N>" placeholders for you to fill in.

Improved Summary & Skills

Drop-in replacements for the section, generated to match recognised industry conventions. Treat them as a strong starting point — adapt to your voice and verify accuracy before using.

Expert insights

5–8 qualitative observations the rules can't catch — for example, "Your most recent role has 8 bullets but only 1 is quantified — promote impact metrics to the top". Use these as a sanity check after applying mechanical fixes.

Privacy & data handling

Your file is not retained. The PDF / DOCX you upload is parsed into text in memory and the temp file is deleted before the response is returned.

What we do store (for paid users only): the extracted text, computed scores, structured model, and AI suggestions — all attached to your account so you can return to the report later.

What we send to the AI provider: only the structured text needed to produce suggestions — never the original file. The provider, model, timeouts, and max-token budget are all configurable in config/config.php.

What we never share: we don't sell, share, or ad-target on your data. Sign-in is via Google Identity Services; we only store your name, email, and avatar URL.

AI Mock Interview

The Mock Interview tier (₹199 / session, with 3-pack and 6-pack bundles) gives you a realtime voice interview with an AI interviewer that adapts to your answers, then scores you across five axes and produces a downloadable PDF report.

How it works. Your browser connects directly to a third-party realtime voice AI Model — none of your voice passes through our servers in real time. The full session is recorded locally and uploaded as a single file at the end so you can replay it on the report.

5 interview types: HR, Behavioural, Technical, Programming (with in-browser Monaco editor + AI code review), and System Design (with a sketching whiteboard).

Modes: Study coaches you after each answer; Real simulates interview pressure with feedback at the end only.

Troubleshooting

The AI keeps talking and doesn't wait for me.

Almost always a mic-pickup problem. Open the live session and watch the green/yellow bar under the mic button — if it doesn't move when you speak, your input is too quiet for voice-activity detection. Move closer, switch to a headset, or check your OS sound settings (Windows: right-click speaker → Sounds → Recording). The session screen also exposes the AI's silence-watchdog: it only fires after you've spoken at least once, so initial thinking pauses won't cause a monologue.

The interrupt button doesn't seem to do anything.

It does — but your browser has buffered a second or two of audio ahead, so the AI's last word may finish playing even after you click. The next time you speak, the AI will respond to what you actually said (not what it would have continued saying). Pressing the spacebar is the keyboard shortcut.

The session disconnected.

We auto-reconnect up to 5 times with exponential backoff (2 s → 4 s → 8 s → 16 s → 32 s). The mic and recording stay live across reconnects. If all attempts fail you'll be sent back to the dashboard — and if you'd barely started, your credit is refunded automatically.

What browsers are supported?

Latest Chrome, Edge, and Firefox on Windows / macOS / Android. Safari is not supported today — it lacks the audio recording APIs (MediaRecorder Opus output) we use for the session replay. On mobile, give the site full-screen mode for fewer interruptions.

Do I need a paid resume tier to do an interview?

No — interviews are a standalone purchase. But if you've uploaded a resume, the AI uses a short summary of it to ground questions (e.g. asking about specific roles or skills you listed). You can also skip the resume entirely.

Can I see what the AI was asked to do?

No — the system prompt is intentionally server-private. It contains your resume summary plus the persona / mode / difficulty configuration. The browser only ever sees the AI's spoken responses and your own transcript.

Privacy: what happens to my recording and transcript?

The combined audio (your mic + the AI voice) and the per-turn text transcript are stored against your account so the report and replay work. We don't share them with any third party. Audio files live under uploads/interviews/<your-user-id>/ with directory-level access denial — they're served only via an auth-gated PHP endpoint scoped to your session. AI evaluation only sees the text transcript and any code submissions, not the audio.

Why is the pre-flight test important?

The pre-flight modal checks browser feature support and grabs your mic before consuming a credit. If anything fails — permission denied, missing API, completely silent mic — your credit isn't spent. Always wait until you see the level bar bouncing before clicking the modal's Start button.

Report FAQ

What's a good total score?

70+ is solid; 85+ is excellent. Below 55 there are usually high-severity fails worth fixing first.

Why is my keyword coverage low when I clearly have those skills?

Two common causes: (1) the industry was misdetected — make your job titles more prominent at the top of the resume. (2) Your Skills section is sparse or buried — list specific tools by name.

My bullet looks fine but the report flagged it. Why?

Check the issue tags above the bullet quote. Common ones: no metric (no number/%/$), passive phrasing, weak verb, too long. The bullet may read well but be missing one of those signals.

Should I use the AI's regenerated Summary verbatim?

No. Treat it as a strong starting point. Adjust to your voice, fix anything inaccurate, and verify it reflects your actual experience.

Why is "Industry not confidently detected" appearing?

Our role detector reads job titles and the top-line keywords. If your resume is generic or the titles are buried, confidence drops below 55% and we fall back to the generic library — which means lower keyword-coverage scores. Make titles prominent in the top 5 lines.

My report says "AI features are disabled" — how do I enable them?

In config/config.php, set AI_PROVIDER to one of openai / anthropic / google / deepseek / ollama and provide an AI_API_KEY. Reload the report — AI sections will populate within a few seconds.

Can I re-score after editing?

The free score is unlimited — re-upload as many times as you want. The detailed paid report (₹99) is per resume; if you want a new detailed report on the edited version, that's a new purchase.

The bullet count looks wrong — half my bullets aren't detected.

The parser detects bullets via list-item markers in your DOCX/PDF. If you used plain paragraphs (no list formatting) instead of bullets, only paragraphs starting with •, -, *, or a numbered prefix are recognised. Convert your Experience entries to actual bulleted lists in your editor.

Why does the report quote my own bullets back to me?

So you can find each problematic line in your resume by Cmd-F / Ctrl-F. Quoting verbatim eliminates ambiguity.

What does "anchor.line_id" mean in the JSON / PDF?

A stable identifier for a specific bullet or section line — e.g. exp-0-b-2 means the third bullet of the first Experience role. Used internally to attach suggestions to the right span.

Ready to score your resume?

The free ATS score takes under a minute. No sign-up required.

Get free ATS score See all plans

How to read your ATS report