Node.js
Read local files, uploads, S3/R2 objects, and other byte sources.
Convert WEBP files to GitHub-Flavored Markdown directly in your browser.
Your WEBP stays on your device. No upload. No account.
Starting local converter
$
npm install @mdgate/image
// Register the callback your app already uses
import { create } from '@mdgate/core';
import { image } from '@mdgate/image';
const toMarkdown = create([
image(yourCallback),
]);
const markdown = await toMarkdown(bytes);
Pure TypeScript, Node.js, Cloudflare Workers, Edge, Browser, No Python, No native addons, No WASM, Zero third-party runtime dependencies
Read local files, uploads, S3/R2 objects, and other byte sources.
Detect images inside Workers. SVG converts locally. Raster images use the callback you register.
Convert SVG locally. Raster images need a vision callback in your application.
Run the same TypeScript converter in compatible Edge runtimes.
SVG text → @mdgate/image → Markdown
JPEG / PNG / … → your vision pipeline
@mdgate/image does not silently send images to an OCR or AI service.
No separate processing service. The converter is TypeScript you can call from Node.js, Cloudflare Workers, Edge runtimes, and browsers.
Detection stays local. Model work happens only through a callback you register.
Recognizes WEBP files from their contents, not only the filename.
Format detection is deterministic. Any model output comes from the callback you supply.
No Python. No native addons. No WASM. No third-party runtime dependencies.
Turn WEBP files into content your agent can actually work with.
@mdgate/image
@mdgate/core
+ @mdgate/image
+ @mdgate/pdf
+ @mdgate/docx
@mdgate/converters
Files you drop on this page stay in the browser. WEBP conversion that needs a model still requires a callback in your application.
Yes. @mdgate/image runs directly in Cloudflare Workers.
No.
No.
GitHub-Flavored Markdown.
Yes. Install @mdgate/image instead of the complete converter set.
Not by itself. Raster formats need a vision callback you provide. SVG converts locally.