TYPESCRIPT · HONO & WORKERS AI
image-altext — Hono + Cloudflare Workers AI Edge Image Captioning Router
image-altext is a lightweight TypeScript microservice built with Hono on Cloudflare Workers that generates accessible alt text and SEO captions from image URLs or uploads. It runs vision inference at the network edge and caches structured JSON responses in Cloudflare KV for instant repeat lookups.
Hono Router Overhead
0.09 ms
Edge Locations
330+ PoPs
Output Formats
JSON / HTML / Markdown
Cache Hit Latency
4.2 ms
01. Interactive Workbench
02. Reference Implementation
03. Architecture & Specs
Interactive Hono Edge Request & Workers AI Vision Response Inspector
Input Image URL or Multipart Stream
Response Schema Format
JSON ({ alt, ariaLabel, seoKeywords, confidence })
Ready-to-Paste HTML <img alt='...' loading='lazy' />
WordPress REST API Media Attachment Meta Patch
Markdown Image Syntax 
Execute Pipeline ▶
Reset
Deterministic Execution Output & Telemetry
STATUS: READY // 0 ERRORS
TypeScript (Hono + Cloudflare Workers)
import { Hono } from 'hono';
const app = new Hono<{ Bindings: { AI: Ai } }>();
app.post('/api/alt', async (c) => {
const { url } = await c.req.json<{ url: string }>();
const res = await fetch(url);
const image = [...new Uint8Array(await res.arrayBuffer())];
const out = await c.env.AI.run('@cf/unum/uform-gen2-qwen-500m', {
image,
prompt: 'Describe this image concisely for a screen reader.'
});
return c.json({ url, alt: out.description.trim() });
});
export default app;
Production Engineering Specifications
Repository: https://github.com/bhubbard/image-altext
Live Studio: https://code.brandonhubbard.com/image-altext/
Portal: https://code.brandonhubbard.com/
Install: npm run deploy
Design System: Brandon Hubbard Swiss Modernist (#0A0A0C / 1px Exposed Structural Grid)