DensityQA LogoDensityQA
AI Work Quality Assurance

Quality Assurance for AI-Assisted Work

DensityQA scans document drafts in 30 seconds to extract unverified claims, missing source citations, and redundant phrasing before client delivery.

Sample Documents

Pilot Program — Low-Friction Entry

Start free, validate with a paid pack, then scale only after the workflow proves out.

Free Evaluation

$0 — 5 Free Audits

Audit 5 document drafts with no credit card required. Includes full claim verification and workslop flags.

Paid Agency Pilot

$149 — 25 Audits

One-time payment • 25 non-expiring audit credits • Includes 30-min strategy review call.

Monthly Plan

$499/mo — Up to 100 Audits

Automated team access for up to 100 monthly document audits with custom agency reporting.

Start with the $149 Agency Pilot
densityqa.com — product demo
Ready to audit

Awaiting first audit

Paste a URL, upload a file, or paste text to begin · Idle

Density score
—
/ 100
Awaiting audit
Est. Time Saved
—
Reader + editorial shortcut

Calculated at ~250 wpm reading + 10 min/1k words editorial review speed.

Fluff
—
Low-value filler
Workslop
0
Clean so far

Claim Checks

Structural rigor of assertions — Supported means verified in-text evidence only.

Claims with direct in-text evidence

Claim checks will populate this table after the audit completes.

Note: DensityQA evaluates structural claim rigor and internal document logic. Claims marked 'Needs Verification' or 'Unsupported' require editorial verification against primary sources.

Workslop findings

Signals that reduce confidence.

0 findings
Flagged workslop cards will appear here.

How Scoring Works

Expand

Density Score

Base 100 minus deductions for workslop, fluff, and unstated assumptions.

Est. Time Saved

Calculated using standard editorial speed (250 wpm reading + 10 min/1k words for manual review).

Understanding Your Density Score

  • 80–100: High Signal

    Supported claims with strong structural grounding, minimal fluff.

  • 50–79: Moderate Workslop

    Contains unstated assumptions or unsupported claims requiring editorial review.

  • 0–49: High Risk

    Heavy corporate fluff, speculative claims, low cognitive density.

Methodology & Limitations

Disclose

Signal density starts at 100 and deducts a fixed amount per flagged finding in code (15 points per speculative / unverified workslop item, 10 points per fluff / redundancy item), then clamps to a 10–100 range. The model extracts findings only — it does not invent the total score.

Claim Checks surface three reader-facing buckets: Supported (direct in-text evidence — "Verified" is reserved for this bucket only), Needs Verification (requires external primary source checking), and Unsupported (no cited evidence). “Unsupported” means missing evidence — not a verdict of factual falsehood.

Est. Time Saved uses ~250 wpm reading plus 10 minutes of editorial review per 1,000 words.

DensityQA checks internal semantic rigor and claim structure. It evaluates logical consistency and unstated assumptions.

Limitations: DensityQA does not perform external fact-checking, web verification, or legal review. Results are heuristic guidance for editorial QA, not a certification of truth.

Audit results load into this dashboard
DensityQA extracts article or document text, then returns a structured cognitive-density audit into the widgets above.

Enterprise Privacy First: Client uploads are processed in memory and never stored, logged, or used to train public AI models.