Generate dynamic 1200×630 Open Graph images for blog posts, products, and landing pages. Pick a template, drop in your title — or describe your content and let AI design it for you.
Template
10 designs
Accent color
#10B981
Use it as a public API — no signup, no credit card
GET /api/og-image takes the same query params: title, subtitle, template, color, and attribution. Free tier is 50 requests / day per IP. Generated images are cached on our CDN for an hour. We don’t log inputs.
For AI Custom: POST to /api/og-image/ai-spec with { prompt }, then render the returned spec via /api/og-image?template=ai-custom&spec=<b64>. AI tier: 5 generations / day per IP.
Frequently asked questions
What is an OG image?
An OG image (Open Graph image) is the preview thumbnail that appears when a link to your page is shared on Twitter, LinkedIn, Slack, Discord, iMessage, Facebook, or any modern app. It's set via a `<meta property="og:image">` tag in your page's HTML and dramatically improves click-through compared to a bare link.
What size should an OG image be?
1200×630 pixels — that's the spec all major platforms (Twitter, Facebook, LinkedIn, Slack, Discord) target as the default. Tinkr's generator outputs exactly that size. Smaller images get upscaled (and look pixelated); larger ones get downscaled by the platform with no benefit.
Is there an API I can call from my app?
Yes. `GET /api/og-image?title=...&subtitle=...&template=default&color=%2310b981` returns a PNG you can drop into a `<meta property="og:image">` tag. Add `&attribution=true` if you want a small "Made with Tinkr" badge in the corner (off by default — no watermark on output). Free tier is 50 requests per day per IP, with `X-RateLimit-*` headers on every response. Generated images are cached on our CDN for an hour, so identical query params don't burn quota.
Is the input I send logged?
No. Tinkr doesn't log titles, subtitles, or any other query parameters. Only an IP-based rate-limit counter is kept (not the inputs themselves). The image is generated on-demand at the edge and never persisted on our side.
How does Tinkr compare to Bannerbear, Placid, and og-image.org?
Cheaper, no signup, no watermark. Bannerbear and Placid start at $49+/month, require an account and API key, and watermark free-tier output. og-image.org is a Vercel demo without a stable templated UI. Tinkr is free for 50 images/day per IP, ships 10 ready-to-use templates including a Canva-style Canvas mode, lets you customize backgrounds / gradients / logos / main images per call, supports AI generation with reference image input, and never adds branding by default — "Made with Tinkr" is opt-in via `?attribution=true`. The whole API is one URL: no auth, no SDK, no per-template configuration.
Is a credit card required to use the API?
No, ever. There's no signup, no email collection, and no credit card required to use the free tier — just hit `GET /api/og-image?title=...` and you'll get a PNG back. Free tier is 50 requests per day per IP. A higher-volume paid tier (with custom templates and longer cache) is on the roadmap, but the free tier will always stay free for individual use.
Can I customize the templates — gradients, colors, logos?
Yes. Every template accepts optional `bgColor` and `textColor` overrides, plus a `logoUrl` for a small brand mark in the top-left. The gradient template additionally takes `gradientTo` and `gradientAngle` so you control the exact gradient. The new `split` template takes a `mainImageUrl` for a large feature image on the right side — drop in your product screenshot, illustration, or hero photo. All fields are optional; leave them off to use the template's default look.
Can I add my own logo to the OG image?
Yes — upload a file from your device or paste an image URL. Click "Upload" next to the Logo field (Canvas mode, or under "Advanced" in Template mode) to pick a JPEG / PNG / WebP / GIF / AVIF up to 5 MB; we store it on our CDN and the URL is permanent so your `og:image` keeps working when shared. Or paste any https:// image URL you already host. The logo renders as a 64×64 badge in the top-left corner of any template. Square images work best.
Can I upload images, or do I have to paste a URL?
Both. Every image field — background (Canvas mode), logo, and main image (Split template) — has an Upload button next to the URL input. Files up to 5 MB (JPEG / PNG / WebP / GIF / AVIF) are stored on our CDN and you get a permanent https URL back. Free tier: 20 uploads / day per IP. Uploads are public (anyone with the URL can fetch them) — that's required for `og:image` meta tags to work when your link is shared, since social platforms fetch the image directly.
What's the 'split' template?
A two-column layout — title and subtitle on the left, a custom image on the right. Set `mainImageUrl` to a screenshot, illustration, or hero photo and the renderer cover-crops it into the right ~42% of the card. Pair it with a `logoUrl` and your accent color and you get a launch-ready product card without opening a design tool.
Can I use my own custom HTML template?
Not yet — but you can already customize the built-in templates extensively (background, text color, gradient stops + angle, logo, main image, accent color), AND the Canvas template lets you drop in any background image and overlay your title (Canva-style). Fully custom HTML/Figma templates are on the roadmap as a paid tier for higher-volume users. The 10 built-in templates plus AI Custom mode cover the vast majority of use cases for free.
What is Canvas mode?
Canvas mode is the closest Tinkr gets to a 'design your own' editor. You paste a background image URL (from Unsplash, your CDN, anywhere https), set a tinted overlay (dark or light at any opacity) for text legibility, and overlay your title and subtitle. An 'Enhance with AI' button next to the title returns 4-5 alternative titles taking different angles (punchy, specific, curiosity, outcome, contrarian). It's free, no signup, no credit card. AI title enhancement: 10/day per IP.
Can I give the AI a reference image?
Yes. In AI Custom mode, upload a reference image (JPEG / PNG / WebP, under 4 MB) — Gemini extracts its dominant palette and overall mood and uses that as a tone anchor for the generated OG. The AI doesn't copy the layout, just matches the aesthetic family. Combine with the 'Additional details' textarea (brand colors, fonts to avoid, audience) for tight control. Free tier still 5 generations/day per IP.
Can I edit the AI-generated design after generation?
Yes. After Gemini returns a spec, click 'Edit' on the spec summary to expose every field — title, subtitle, background color, text color, accent color, layout (centered / split / left-aligned / magazine / minimal), font style (sans / serif / mono / display), decorative element, title size, subtitle size. Edits update the live preview in real time and don't burn extra AI quota. Use this to adjust palette to your exact brand or to tweak title length without re-running the AI call.
How does AI title enhancement work?
Type a title in Canvas mode and click 'Enhance with AI'. Gemini returns 4-5 alternatives, each taking a meaningfully different angle: punchy (terse, verb-led), specific (concrete numbers), curiosity (opens a loop), outcome (focuses on the reader's gain), or contrarian (challenges an assumption). Each suggestion includes a one-sentence reason explaining why that framing tends to work on social shares. Click any suggestion to use it. Free 10/day per IP, no signup, no credit card.
Does the image have a Tinkr watermark?
No, never by default. The generated PNG looks like your own work — no "Generated with Tinkr" text, no logo, no domain. If you'd like to credit Tinkr (we appreciate it!), turn on the "Add 'Made with Tinkr' badge" toggle on the page or pass `&attribution=true` to the API. The badge is a tiny ~50% opacity pill in the bottom-right corner.
Why does my image look pixelated?
It's not — the source PNG is 1200×630, which is the full social-share resolution. Some platforms (Discord, iMessage) downscale OG images aggressively in chat previews, which can introduce mild blur. The image is sharp at native size; if you embed it directly in a webpage, render it at 1200×630 or smaller without upscaling.
Is this free?
Yes. The free API tier gives every IP address 50 image generations per 24 hours. Rate-limit headers (`X-RateLimit-Limit`, `X-RateLimit-Remaining`, `X-RateLimit-Reset`) are included on every response so you can see how many calls you have left. A higher-volume paid tier with custom templates is on the roadmap.
What's the AI Custom mode?
Describe your content in plain English (a blog post topic, product launch, video, etc.) and Gemini picks the colors, layout, typography, and decorative elements that match the vibe — then we render it into a 1200×630 PNG. No image-generation API, no cost-per-image AI billing — just text → structured design spec → JSX render. Free 5 generations per day per IP, no signup, no credit card.
How is AI Custom different from picking a template?
Templates are great when one of the 9 built-in styles fits — pick default/dark/gradient/code/quote/minimal/bold-serif/pastel/split, drop in your title, optionally customize the background and logo, done. AI Custom is great when you want a unique design tuned to YOUR specific content — a calm pastel for a barista video, a bold gradient for a SaaS launch, an editorial serif for a long-form essay. Both produce a 1200×630 PNG via the same `/api/og-image` endpoint.