Quick Start • Features • Examples • Contributing

LandingAudit scans a folder of local HTML landing pages and produces actionable navigation and conversion‑flow insights. It is aimed at SaaS founders, product marketers, and growth teams who want to improve conversion rates without manual UX audits.

$ python -m landingaudit.main --input ./landing_pages
Scanned 57 pages, found 12 issues (3 critical, 5 high, 4 medium).
Reports written to ./reports/audit_report.csv and ./reports/audit_report.json

Most SaaS landing pages have broken navigation flows that hurt conversions, but founders lack systematic ways to identify these issues. Manual analysis is time‑consuming and requires UX expertise. Companies need data‑driven insights about what's working across their industry to benchmark their own pages.

Feature	Description
Page Crawl & Data Extraction	Recursively walks the input directory, loads each HTML file, and extracts anchors, forms, and CTA‑related text using only the Python standard library.
Navigation Flow Detection	Builds a directed graph from extracted links to surface orphaned pages, circular navigation, missing CTAs, and excessive redirect chains.
Conversion Issue Reporting	Evaluates forms and CTA prominence, flags missing required attributes, incorrect input types, excessive form fields, and generic CTA text.
Offline‑First Analysis	Runs entirely offline with zero external API calls; all parsing, graph building, and scoring use the standard library, ensuring data never leaves the machine.
Configurable Sensitivity	Adjust the minimum confidence threshold via the LF_MIN_CONFIDENCE environment variable (or use --input/--output flags for paths) to tune how aggressively issues are reported.
Structured Report Generation	Emits both CSV (RFC 4180) and JSON (RFC 8259) files that downstream tools can ingest without schema validation errors.

Quick Start

1. Clone the repository: git clone https://github.com/m2ai-portfolio/landingaudit.git
2. Change into the project directory: cd landingaudit
3. Ensure Python 3.11+ is installed (python --version).
4. (Optional) Set environment variables: export LF_DATA_DIR=./landing_pages, export LF_OUTPUT_DIR=./reports, export LF_MIN_CONFIDENCE=0.7.
5. Place your HTML landing pages in the landing_pages folder.
6. Run the audit: python -m landingaudit.main

Examples

Basic audit of a sample landing page set

$ python -m landingaudit.main --input ./demo_pages
Scanned 23 pages.
Issues found:
- ./demo_pages/pricing.html: missing_cta (severity 8) – Add a prominent CTA button above the fold.
- ./demo_pages/features.html: broken_navigation (severity 6) – Fix orphaned anchor pointing to /non‑existent.
Reports written to ./reports/audit_report.csv and ./reports/audit_report.json

Audit with a higher confidence threshold to surface only critical issues

$ LF_MIN_CONFIDENCE=0.9 python -m landingaudit.main
Scanned 23 pages.
Issues found (confidence ≥ 0.9):
- ./demo_pages/checkout.html: missing_cta (severity 9) – Primary CTA not detectable.
Reports written to ./reports/audit_report.csv and ./reports/audit_report.json

Specify a custom output directory

$ python -m landingaudit.main --output ./custom_reports
Scanned 23 pages.
Issues found: 12.
CSV report: ./custom_reports/audit_report.csv
JSON report: ./custom_reports/audit_report.json

File Structure

landingaudit/
├─ landingaudit/
│   ├─ __init__.py
│   ├─ main.py
│   ├─ core/
│   │   ├─ __init__.py
│   │   ├─ crawler.py      # HTML traversal and extraction
│   │   ├─ navigator.py    # Graph building and issue detection
│   │   └─ reporter.py     # CSV/JSON report generation
│   └─ data/
│       ├─ __init__.py
│       └─ models.py       # Issue and AuditResult dataclasses
├─ landing_pages/          # Place your HTML files here
├─ tests/                  # Unit tests + HTML fixtures
├─ reports/                # Generated audit reports (CSV/JSON)
├─ assets/
│   └─ infographic.png    # Banner image used in README
├─ .gitignore
├─ LICENSE
└─ README.md

Tech Stack

Technology	Purpose
Python 3.11+	Runtime language
pathlib	Filesystem traversal
html.parser	HTML parsing
re	Pattern matching for CTA detection
json	Structured report output
csv	Tabular report output
typing	Type hints for clarity
dataclasses	Lightweight data models

Contributing

Fork the repository, create a feature branch, make your changes, ensure existing tests pass (python -m unittest discover tests), and submit a pull request. Please follow the existing code style and add unit tests for new functionality.

License

MIT

Author

Matthew Snow -- M2AI | @m2ai-portfolio