Content Accessibility Utility on AWS

Digital content stakeholders across industries aim to streamline how they meet accessibility compliance standards efficiently. The Content Accessibility Utility on AWS offers a comprehensive solution for modernizing web content accessibility with state-of-the-art generative AI models, powered by Amazon Bedrock. It automatically audits and remediates WCAG 2.1 and 2.2 accessibility compliance issues, and offers a Python CLI, a Python API, and a deployable event-driven pipeline. Capabilities include batch processing for large volumes of content, usage tracking for cost management, and will continue to expand to support other content types and modalities.

📚 Documentation

Full documentation is published at awslabs.github.io/content-accessibility-utility-on-aws.

Start there for guides on every way to use the solution:

• CLI Guide — install and run from your terminal
• API Integration Guide — embed it in your own Python app
• Deployable Pipeline Guide — process documents automatically at scale
• Parameter Reference, Remediation, Rendered Audit & Agent, Architecture, and more.

The documentation source lives in docs/.

Features

• Convert PDF documents to accessible HTML, preserving layout and visual appearance
• Extract and embed images
• Audit HTML for WCAG 2.1 and 2.2 accessibility compliance
• Remediate common accessibility issues using Amazon Bedrock models, including advanced table remediation strategies
• Optional browser-backed (rendered) audit that detects computed-style and interactive issues static HTML analysis cannot see (e.g. focus visibility), using a real headless browser and axe-core
• Optional accessibility agent (Strands) that drives a render → fix → verify loop, confirming each fix actually renders correctly before marking it resolved
• Optional internationalization ([i18n]) that translates worked-on content into one or more target languages via Amazon Bedrock, preserving markup and screen-reader-announced attributes and optionally emitting a single multilingual document with an accessible language selector
• Single-page and multi-page output formats
• Batch processing for large-scale document processing
• Detailed usage tracking for BDA pages and Bedrock tokens, plus cost analysis
• Streamlit sample web interface with usage visualization

Installation

# From PyPI
pip install content-accessibility-utility-on-aws

# From source
pip install .

The core install is static-only (no browser). The rendered audit, the agent, and translation are opt-in extras so the base footprint stays small:

pip install "content-accessibility-utility-on-aws[rendered]"   # browser-backed audit
pip install "content-accessibility-utility-on-aws[agent]"      # render -> fix -> verify loop
pip install "content-accessibility-utility-on-aws[i18n]"       # translation + multilingual output

The [rendered] and [agent] extras use a headless browser via Playwright; after installing either, run playwright install chromium once. The core package never imports the browser or agent stack, so static-only workflows are unaffected. See the Installation notes for details.

Quickstart

The audit path needs only a pip install and an HTML file — no AWS account required:

pip install content-accessibility-utility-on-aws
content-accessibility-utility-on-aws audit -i page.html -o report.html -f html

Open report.html for a human-readable accessibility report. For the full convert → audit → remediate workflow on a PDF (which uses Amazon Bedrock), see the CLI quickstart.

Pick the way you want to use it

Prerequisites

• Python 3.11+
• An AWS account with access to Amazon Bedrock models and, for the PDF conversion path, Bedrock Data Automation (BDA).
• An S3 bucket for BDA file uploads during PDF conversion.

The PDF path reads BDA_S3_BUCKET and BDA_PROJECT_ARN from the environment (or accepts --s3-bucket / --bda-project-arn). Auditing and remediating existing HTML only needs Bedrock model access. Full setup — creating the S3 bucket and BDA project, IAM, and configuration files — is in the CLI Guide prerequisites.

Architecture

The package consists of four modules — PDF2HTML, Audit, Remediate, and Batch — plus an optional browser-backed agent layer.

See Architecture & Core Packages for the per-module breakdown and diagrams.

Requirements

• Python 3.11+
• AWS credentials for Bedrock Data Automation and Bedrock models
• Appropriate IAM permissions for S3 and BDA services

Configure AWS credentials with aws configure, environment variables (AWS_ACCESS_KEY_ID / AWS_SECRET_ACCESS_KEY), or a named profile via --profile.

License

Apache-2.0 License. See LICENSE for details.

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for details on how to contribute to this project.