Add AI media generation
to your product

Two layers of control. First, you set AI brand guidelines — style description, color palette, tone keywords, and elements to avoid. Those instructions shape supported workflows so generated visuals and copy stay aligned with your brand. Second, image outputs can add logo and layout overlays in the branding editor after generation. Your users can't bypass either layer.

ImageLayer runs production generations on Google Vertex AI: Gemini image models for visuals, Gemini text models for copy, Veo 3.1 Lite/Fast/Quality for video, and Gemini TTS for voice. We also use Gemma 4 on Cloudflare Workers AI for moderation and prompt enrichment. You can try a limited public preview without signing in, then switch the playground to your live workspace after authentication.

Images, video, voiceovers, podcast-style audio, and text. Plan controls decide which modes are available to your users. Most teams start with image workflows, then unlock video, voice, and text as they expand the experience.

Pre-built forms for common creative outputs — stat highlights, quote cards, blog promos, announcements, and infographics. Instead of writing a prompt from scratch, users pick a template, fill in a few fields, and generate. Depending on your plan, you can create up to 3, 10, or 100 custom templates for your specific use cases.

Yes. Built-in platform presets cover LinkedIn (post, carousel, banner), Instagram (post, story), X/Twitter, Facebook (post, story), Google Business, WhatsApp Status, Telegram Post, Pinterest Pin, TikTok Cover, YouTube Thumbnail, blog headers, email banners, and custom dimensions. Each preset sets the correct dimensions and optimizes the composition for that platform. Choose Custom if you need a different aspect ratio.

Add one <image-layer> HTML tag — that's the entire integration. It's a native Web Component, so it works with any framework: React, Vue, Angular, Svelte, Next.js, or plain HTML. The same widget can expose image, video, voice, and text workflows. Most teams go from sign-up to live widget in under 15 minutes.

Yes. The widget is a native Web Component — the same standard browsers support natively. It works out of the box with React, Vue, Next.js, Nuxt, Angular, Svelte, SvelteKit, Astro, Remix, and any other framework. It also drops into WordPress, Webflow, and other no-code tools. No npm package required — you can load it from our CDN with a single script tag.

Yes. You can control brand rules, content templates, enabled modes, and other dashboard settings for each workspace. Your dashboard also shows who created what, when, and how many credits were used.

You'll get a notification before you run out. You can upgrade your plan anytime with no downtime. Enterprise customers get volume pricing, custom credit pools, and SLAs.

We use Google Veo 3.1 on Vertex AI for video generation, with Lite, Fast, and higher-quality tiers depending on the workflow. It produces 720p or 1080p (Pro+) videos up to 8 seconds long. Lite starts at 7 credits, Fast typically costs 13-25 credits depending on duration and resolution, and the highest-quality tier costs more.

ImageLayer offers 8 distinct AI voices powered by Google Gemini TTS — from firm and professional to warm and conversational. Choose between Flash Lite (lowest-cost single-speaker), Flash (faster general-purpose), and Pro (richer intonation) models. Output is MP3 or WAV format.

Yes. Pro plan users can generate two-host discussion segments and interview-format audio with multi-speaker TTS models like Gemini 2.5 Flash TTS and Gemini 2.5 Pro TTS. The system automatically pairs complementary voices for natural-sounding conversations. Write your script in conversational style and the TTS handles speaker alternation.

Connections use HTTPS/TLS. API keys are SHA-256 hashed and never stored in plaintext, passwords are stored as bcrypt hashes, and temporary generated media is cleaned up on a rolling TTL. We require email verification before live API use, apply rate limits and tenant scoping, and do not use your prompts or generated assets to train our own models. See our Security page for full details.