ALT-Kai · AI Alt-Text Generator
Accessibility AI · Multi-Step Gemini Agents
AI alt-text generator for technical and research content. Ingests PDFs, Excel sheets, or batches of images and outputs WCAG 2.2 + BenTech-compliant alt-text using a multi-step Gemini agent pipeline.
The Brief
Problem
A research publishing client had thousands of PDFs and figure assets shipping to print without compliant alt-text — a regulatory risk and a barrier for visually-impaired readers.
The Architecture
Decision
Multi-step Gemini agent chain: ingest & decompose → context retrieval → caption draft → WCAG 2.2 rule check → BenTech compliance check → confidence scoring. Output to publishable format with manual override for low-confidence captions.
The Outcome
Result
Compliance backlog cleared. Every new asset auto-tagged on upload. Editors review only flagged cases.
How it actually works in production.
Ingest
PDF batch
research papers
Excel sheet
figure manifest
Image batch
standalone assets
Decompose
extract images + ctx
Multi-Step Gemini Agents
Context retrieval
surrounding text
Caption draft
first pass
WCAG 2.2 check
rule engine
BenTech check
compliance rules
Confidence
low → review
Output
Compliant alt-text
WCAG 2.2 + BenTech
Publishable export
embed in source doc
Animated · Built in code · No GIFs
See ALT-Kai · AI Alt-Text Generator in action.
Full ALT-Kai walkthrough — PDF upload → multi-step Gemini pipeline → publishable alt-text
ALT-Kai live homepage — landing page
Pipeline status — file upload → save → extract text + images, AI figure mapping starts
Multi-step Gemini pipeline running — generating alt-text 3 of 8 figures
Mid-pipeline — image extraction + WCAG / BenTech compliance check stages
Processing complete — generated alt-text table with figure previews + short and detailed captions for scientific protein structures
Final output — JSON + CSV export ready for publication
Stack