AI-powered accessibility toolkit. WCAG audits, code fix generation, alt text evaluation, and structured reports — powered by local AI.
Built by Like One Foundation — because accessibility shouldn't require a six-figure budget.
pip install techaccess
playwright install chromium# Full WCAG audit with score
techaccess audit https://example.com
# AI-enhanced audit (semantic analysis via Ollama)
techaccess audit https://example.com --ai
# Generate code fixes for accessibility issues
techaccess fix https://example.com
# Evaluate alt text quality
techaccess evaluate https://example.com
# ARIA accessibility tree snapshot
techaccess snapshot https://example.com
# Contrast ratio check
techaccess contrast https://example.com# JSON, Markdown, SARIF (GitHub Code Scanning), or HTML report
techaccess audit https://example.com -f html -o report.html
techaccess audit https://example.com -f sarif -o results.sarif
# CI mode — exit code 1 on serious+ issues
techaccess audit https://example.com --fail-on serious# Add to Claude
claude mcp add techaccess -- python /path/to/server.pyTools: access_audit, access_fix, access_evaluate, access_snapshot, access_contrast, access_alttext
TechAccess combines three layers:
- axe-core — industry-standard WCAG 2.0/2.1/2.2 rule engine
- Playwright ARIA — captures the full accessibility tree for semantic analysis
- AI layer — local models (Ollama) or Claude API for:
- Semantic ARIA analysis (landmark structure, heading hierarchy, focus order)
- Alt text generation and evaluation (llava for vision)
- Code remediation with before/after diffs
- Contextual fix suggestions
0-100 score with letter grade (A+ through F). Weighted by severity: critical issues count 4x more than minor ones.
- JSON — machine-readable, CI/CD integration
- Markdown — human-readable summaries
- SARIF — GitHub Code Scanning compatible
- HTML — dark-themed dashboard with charts and fix previews
- Python 3.10+
- Chromium (installed via Playwright)
- Optional: Ollama for local AI features
MIT. Free forever.