OpenGraph.io plugin

• Credits, not side effects. Every operation is a read, but each request bills credits: 1 for a base call or cache hit, 10 for full_render or use_proxy, 20 for use_premium or a screenshot, 30 for use_superior. query.run bills a fixed 100 credits on the default nano model and 200 on mini, whatever proxy or render options are set, so one uncached query uses up a free account’s month. On v3.0 auto_proxy, auto_render and retry default to true, so OpenGraph.io may pick a proxy tier for a hard domain and bill that tier.
• Rate limits are concurrent. Plans allow 1 (free) to 100 requests in flight at once. Past that the API answers 429; the plugin retries up to three times with exponential backoff and honors Retry-After when sent.
• Enum-typed options. screenshot.capture takes format (jpeg, png, webp) and dimensions (xs, sm, md, lg); oembed.get takes orientation (vertical, horizontal), theme (light, dark, auto) and template (default, compact, horizontal); query.run takes modelSize (nano, mini); extract.run and markdown.get take ai_sanitize_mode (sanitize, warn, block); an extract.run selector’s type is text or attr; a brand.analyze question’s type is one of string, boolean, number, string[] or object, which name the answer’s type. Pass them all as strings, e.g. type: 'string[]'.
• Screenshot links expire. screenshot.capture returns a CDN URL that dies after 24 hours; download the image if you need it longer.
• Markdown always comes back as the JSON envelope, never a bare text body, so chunking, structure and usage metadata parse uniformly.
• Brand analysis is slow and partial. brand.analyze reads several pages of the site and can take tens of seconds; sections that fail come back null rather than failing the call. Requests with sections or questions POST a JSON body, plain ones GET.
• Webhooks. OpenGraph.io sends none, so the plugin receives none.

• 8 typed API operations

​

Setup

npm install corsair @corsair-dev/opengraph better-sqlite3
npm install --save-dev @types/better-sqlite3

yarn add corsair @corsair-dev/opengraph better-sqlite3
yarn add --dev @types/better-sqlite3

pnpm install corsair @corsair-dev/opengraph better-sqlite3
pnpm install --save-dev @types/better-sqlite3

bun add corsair @corsair-dev/opengraph better-sqlite3
bun add --dev @types/better-sqlite3

import Database from 'better-sqlite3';
import { createCorsair } from 'corsair';
import { opengraph } from '@corsair-dev/opengraph';

export const corsair = createCorsair({
	plugins: [
		opengraph(),
	],
	database: new Database('corsair.db'),
	kek: process.env.CORSAIR_KEK!,
	multiTenancy: true,
	hub: {
		projectApiKey: process.env.CORSAIR_API_KEY!,
		signingSecret: process.env.CORSAIR_SIGNING_SECRET!,
	},
});

opengraph()

const { connectUrl } = await corsair.manage.connect.createLink({
	plugin: 'opengraph',
	tenantId: 'acme',
});
// redirect the user's browser to connectUrl

​

Example API calls

const tenant = corsair.withTenant('acme');
await tenant.opengraph.api.site.get({ url: 'https://github.com' });

const tenant = corsair.withTenant('acme');
await tenant.opengraph.api.query.run({ url: 'https://en.wikipedia.org/wiki/Web_scraping', query: 'What legal issues surround web scraping?' });

​

What’s next