EchoScan

Content Distinctiveness & Information-Gain Auditor Built by Brewcontent.ai

Marketing teams generate more AI-assisted content than ever, and it's converging on the same phrasing, structure, and claims. Research on the 2026 content landscape consistently points to the same root problem: the content gap is no longer a lack of keywords — it's a lack of information gain, unique data or perspective that consensus-trained models can't produce, buried under AI-tell cliches that hurt both reader trust and AI-search visibility.

EchoScan is a small, focused CLI that scores any piece of content across three dimensions before you publish it:

1. Genericness — detects marketing/AI-tell cliches ("in today's fast-paced world," "unlock the full potential," "game changer") and filler-sentence density.
2. Information Gain — measures concrete evidence (numbers, dates, named sources, citations) against vague, unsupported claims ("studies show," "experts agree").
3. Competitor Overlap (optional) — TF-IDF cosine similarity against competitor content you supply locally, with shared-term breakdown.

These roll up into a single Distinctiveness Score (0–100) with a verdict and concrete, line-level suggestions — runnable locally, in CI, or as a pre-publish gate.

Install

pip install -e .

Requires Python 3.9+.

Usage

echoscan analyze path/to/content.txt

With competitor comparison:

echoscan analyze path/to/content.txt \
  --competitor path/to/competitor_a.txt \
  --competitor path/to/competitor_b.txt

Export the full report as JSON (for CI gates or dashboards):

echoscan analyze path/to/content.txt --json report.json --quiet

Try the bundled examples:

echoscan analyze examples/generic_sample.txt
echoscan analyze examples/distinctive_sample.txt --competitor examples/competitor_sample.txt

Example output

EchoScan — Content Distinctiveness & Information-Gain Auditor
Built by Brewcontent.ai

generic_sample.txt  (142 words)
Distinctiveness Score: 18.4/100 — Highly generic — rewrite recommended before publishing.

Suggestions:
  → Cut or replace overused phrases (top offenders: "in today's fast-paced world",
    "game changer", "unlock the full potential")...
  → Add concrete evidence: original data, named sources, dates, or percentages...

Library usage

from echoscan import analyze

report = analyze(open("post.md").read(), source="post.md")
print(report.distinctiveness_score, report.verdict)

Scoring methodology

All scoring is deterministic, local, and offline — no external API calls, no data leaves your machine:

• Genericness uses a curated cliche/filler-phrase library (echoscan/patterns.py) plus sentence-level filler detection.
• Information Gain uses regex-based evidence markers (percentages, currency, years, citations, sample sizes) and a lightweight named-entity heuristic, offset by vague-claim phrase detection.
• Overlap uses scikit-learn's TF-IDF vectorizer + cosine similarity against competitor text you provide directly (no scraping).

The pattern libraries in patterns.py are intentionally the easiest place to extend accuracy over time — add phrases relevant to your industry as you see them recur.

See docs/ARCHITECTURE.md for the full design rationale, trade-offs, and extension points.

Development

pip install -e ".[dev]"
pytest
ruff check src tests

CI runs the full test suite across Python 3.9–3.12 on every push (.github/workflows/ci.yml).

Roadmap ideas

• Pluggable pattern packs per vertical (SaaS, fintech, healthcare)
• Optional embedding-based overlap scoring (vs. TF-IDF) for semantic similarity
• GitHub Action for automated pre-publish gating
• Batch/directory mode for auditing an entire content library at once

License

MIT — see LICENSE.

---

Built by Brewcontent.ai