Changelog
What's new in ImageLayer
New features, improvements, and fixes — shipped continuously.
Updated Google Model Catalog
ImageLayer now uses the current Gemini 3 image catalog, adds Gemini 3.5 Flash for text generation, and lets Gemini image requests choose 1K, 2K, or 4K output quality with matching credit pricing.
Campaign and Promo Bundles
ImageLayer can now generate coordinated campaign bundles from product details or B2B campaign briefs. Users can review grouped outputs, submit selected assets back to the host app, and see campaign items grouped in history.
Security Hardening for CI, Sessions, and Widget Embeds
ImageLayer now has stronger safeguards against supply-chain and token-exposure risks, including tighter CI controls, signed session creation, private generated media URLs, and safer widget handling for embedded history and settings.
White-Label Smart History and Settings
B2B customers can now embed ImageLayer history and sub-account settings inside their own product. History search supports media-type filters and can use indexed visual context from generated images and videos, so teams can find past outputs by what they contain, not just by the original prompt.
OpenAI Models Across Image, Text, Voice, and Video
ImageLayer now supports OpenAI models alongside Gemini and Vertex AI: GPT Image 2 for premium image generation, GPT-5.4 mini for text, Sora 2 for video, OpenAI voices for TTS, and GPT-powered asset analysis. Teams can choose the model family that best fits quality, cost, and creative direction without changing their workflow.
Find Similar Past Generations
A new POST /v1/generate/similar endpoint surfaces the org’s past generations whose prompts are semantically closest to a current prompt, powered by OpenAI embeddings. Foundation for upcoming widget history smart features.
OpenAI Sora 2 Premium Video
Sora 2 is now available alongside Veo 3.1 as a premium video model. Pick it for marketing-grade output with explicit camera direction; the same brand guidelines, queueing, and billing flow apply.
OpenAI Voices for Voiceovers
OpenAI TTS now ships alongside Gemini TTS with six new voices — Alloy, Echo, Fable, Onyx, Nova, Shimmer — for narration, ads, and creative voice direction. Pick the voice that fits your brand without leaving the widget.
Higher-Fidelity Asset Analysis
Asset analysis (AI summary, alt text, brand-element observations, compliance notes) now uses OpenAI GPT-5.4 Vision when available, falling back automatically to the prior provider — typically catches more brand-relevant detail in logos, screenshots, and marketing materials.
GPT-5.4 mini for Text Generation
You can now pick OpenAI GPT-5.4 mini as an alternative text-generation model alongside Gemini. Lays the groundwork for schema-guaranteed JSON outputs for content type fields in an upcoming release.
Dual-Layer Prompt Moderation
Prompt safety checks now run on two independent providers in parallel — adding redundancy and reducing the impact of any single moderation provider outage on widget end users.
OpenAI GPT Image 2 Available Alongside Gemini
You can now pick OpenAI GPT Image 2 from the model selector for higher fidelity output. The new option flows through the same brand guidelines, references, queueing, and billing path as the existing Gemini models, and is automatically gated by per-minute and hourly safeguards so it stays reliable under load.
Gemini 3.1 Flash TTS in Audio Generation
Audio generation now includes Gemini 3.1 Flash TTS, giving you a newer speech option with more natural delivery and stronger expressive control for voiceovers and spoken content.
Expanded AI Model Catalog and Pricing Alignment
The available AI model catalog now includes Gemini 3 Pro Image, Gemini 3.1 Pro for text generation, and faster Veo 3.1 video tiers. Model labels and pricing were aligned across the widget, dashboard, and backend so the options you see match current supported behavior.
Refine and Re-Roll Past Generations
Generation History now supports follow-up refinements and same-seed re-rolls, so you can improve a result or generate a close variation without starting from scratch.
Manual Logo Upload and Reference Gallery
You can now upload a logo manually and manage a richer reference-image gallery, giving the widget stronger brand context even when website-based extraction is not enough.
Clearer Brand and Safety Settings
Brand Guidelines and dashboard settings were reorganized into a simpler wizard and tabbed editor, with moderation controls moved under Safety for easier setup.
History Page No Longer Bounces to Sign-In
Opening Generation History from the dashboard now loads correctly for cookie-based sign-in instead of looping between the page and the login screen. A safety guard also prevents similar redirect loops in the future.
Insights Card Loads Reliably
Fixed a regression that returned a 500 from the dashboard Insights endpoint, so first-attempt acceptance, regenerations, and refine-usage signals now render again on the overview.
Playground Page Stops Crashing on Navigation
The interactive playground used to fail on certain in-app navigations with a "cannot read properties of undefined" error. It now falls back to a safe default and keeps working regardless of how you arrived on the page.
Website-Based Brand Autofill
You can now paste a public website URL into Brand Guidelines settings and generate an editable draft from the site’s branding signals before saving.
Temporal Patterns in Dashboards
Admin and organization dashboards now show which weekdays and hours attract the most generation traffic, plus where slower latency clusters, so peak load windows are easier to spot.
Industry Preset Starter Kits
You can now apply ready-made bundles of content types, prompt suggestions, and platform presets tailored to your business, so your widget starts with relevant defaults instead of a blank setup.
AI Summary in Overview
Paid organizations now see an AI-generated summary at the top of Overview with the biggest usage changes, reliability signals, and concentration trends. Enterprise workspaces also get more actionable recommendations.
Admin Failure Explorer
Superadmins can now inspect blocked and failed generation requests with prompt snippets, moderation reasons, provider request IDs, and input context directly from the admin overview and per-organization analytics.
Deeper Admin Org Analytics
Admin analytics now include model health, request-pipeline status counts, moderation signals, credit-pack activity, and adoption metrics for sketches, references, and composites.
Full Generation History
A dedicated History tab in the widget lets you search, filter by type, and browse all past generations — including text, audio, and video — with infinite scroll and bulk clear.
Background Generation Recovery
If you start a generation and leave the page, the widget automatically recovers and shows the result when you return.
History Toggle in Dashboard
Organizations can now enable or disable generation history for their widget users from the Customization settings.
Mock Provider for Load Testing
A new mock provider mode lets you stress-test queue and infrastructure capacity without consuming Vertex AI credits.
Generate Studio
Use the full widget directly from the dashboard — no embedding needed. Great for testing your setup before integrating, or for teams that want to generate content without a website.
Feature Request Button
Request new features directly from your dashboard. Every request goes straight to the team — we read and respond to all of them.
Default Model Upgrade
The default image model is now Gemini 3.1 Flash Image, offering better aspect ratio support and improved reference image handling at the same credit cost.
Accurate Generation Timer
The estimated generation countdown now reflects real-world processing times more accurately, so you know when to expect your results.
Admin Panel Polish
Improved form styling across the admin panel, searchable organization picker, and a dedicated admin test generation page for end-to-end pipeline verification.
Model Usage Analytics
See which AI models your users are choosing. The dashboard now shows a model breakdown donut chart and per-model performance metrics including latency and failure rates.
History Item Deletion
Delete individual generations from your history by hovering over a thumbnail and clicking the X button. Removal is instant and frees up associated resources.
Richer Analytics & Reliability Insights
Organization and superadmin dashboards now show failure rates, latency trends, output-mode performance, top errors, and queue health so issues are easier to spot and diagnose.
Google Sign-In for Organizations
Organizations can now sign up and sign in with Google in one click. Existing accounts are linked securely, so teams keep the same workspace and onboarding flow.
Vertex AI Catalog and Validation Sync
Image, video, audio, and text generation now use a synchronized Vertex AI model catalog across the widget, API, pricing, and docs, so supported options and credit behavior stay consistent end to end.
Audio Generation & Text-to-Speech
Generate high-quality speech and voiceovers directly from the widget using Google Gemini TTS. Choose from 8 distinct voices, select MP3 or WAV format, and control duration up to 5 minutes.
Podcast Snippet Generation
Create two-host podcast-style audio from any topic — the AI writes the script and synthesizes both voices automatically. Perfect for social media content and LinkedIn posts.
Video Voiceover
Add AI-generated voiceovers to any completed video with one click. Audio is merged with the video using ffmpeg so you get a single downloadable file.
Video Generation (Veo 3.1 Lite)
Generate short videos from text prompts or source images using Google Veo 3.1 Lite. Supports 4–8 second clips at up to 1080p, with brand guidelines automatically applied.
Nano Banana 2 Image Model
Added Google's Nano Banana 2 (Gemini 3.1 Flash Image) as a new generation model — delivers character-consistent images with accurate text rendering at just 5 credits.
AI Content Moderation
Pro plan admins can now enable AI-powered content moderation to automatically filter inappropriate images and validate prompts. Configurable categories, blocked words, and enforcement modes available in dashboard settings.
Prompt Enhancement & Alt-Text Generation
AI now enriches short prompts into detailed generation instructions and auto-generates alt text for accessibility and SEO — powered by Gemma 4 on Cloudflare Workers AI.
Video-Aware Widget UX
The generation widget now fully supports video mode with a dedicated tab, video-specific loading messages, native video player controls, and duration/resolution settings.
Refreshed Loading Experience
Replaced the animated logo during generation with a clean ring loader animation and expanded the loading message pool to 20+ phrases so they never repeat during a single generation.
Dedicated Video Queue & Credit System
Video generation runs on its own BullMQ queue with extended lock durations. Credit costs are dynamically calculated based on model, duration, and resolution.
API Key Creation Modal
Creating API keys now opens a focused modal dialog instead of an inline form, making the flow clearer and ensuring you don't miss copying your new key.
Contextual Tooltips Across Dashboard
Added helpful tooltips to stat cards, table columns, chart legends, and settings tabs so you can understand metrics and controls without leaving the page.
Mobile-Friendly Tooltips
Tooltips now only appear on devices with a mouse or trackpad, preventing confusing tap interactions on phones and tablets.
Preset Guides & Case Studies
Published in-depth guides for every content-type preset — announcements, quote cards, testimonials, product photos, infographics, and more. Added brand showcase case studies demonstrating real-world usage.
Prompt Engine Improvements
Refactored the prompt pipeline for better consistency across generation modes and content types.
Blog Launch
Launched the blog with guides, tutorials, and product updates. Includes full-text search and tag-based filtering.
Performance & SEO Optimization
Lighthouse audit and performance pass across all pages — improved load times, Core Web Vitals, and search engine indexing.
Framework Integration Guides
Added dedicated integration pages for React, Vue, Next.js, Angular, Svelte, and Astro with copy-paste code snippets.
Brand Guidelines
Configure brand colors, tone of voice, and reference images in the dashboard. The AI uses these guidelines to produce visuals that match your brand identity automatically.
Content-Type Templates
Structured content types replace freeform aspect-ratio selection. Choose from LinkedIn posts, Instagram stories, event covers, product shots, and more — each with optimized format presets and prompt defaults.
Analytics & Monitoring
Added usage analytics tracking across the platform for better visibility into generation patterns and performance.
Interactive API Documentation
Published OpenAPI-powered API reference with Scalar — browse endpoints, try requests, and view schemas directly in the browser.
Platform-Ready APIs
Added sub-account management, per-account quota endpoints, and usage-by-account reporting for platform integrators.
Free Plan Hardening
Email verification required before API key creation. Added multiple layers of abuse prevention to keep the free tier fair for everyone.
Security & Trust Page
Published a dedicated security page with encryption practices, data handling policies, and links to independent verification tools.
Reliable Generation Queue
Moved image generation to a background job queue with automatic retries, backoff, and concurrency control. Eliminates dropped requests under load.
LemonSqueezy Payments
Integrated LemonSqueezy as merchant of record for subscriptions and credit packs. Added webhook handling, customer portal for self-service billing, and free tier registration.
Credit-Based Pricing
Introduced tiered plans with credit limits and purchasable credit packs. Usage tracked at the organization level with real-time quota enforcement.
Test Suite & Node.js 24
Added unit and integration tests for core packages. Upgraded runtime to Node.js 24 LTS.
ImageLayer Launch
Initial release of the embeddable AI image generation widget. Generate images from text prompts or rough sketches, with automatic brand overlays — logos, color palettes, and layout frames applied to every output.
Organization Dashboard
Real-time dashboard with usage analytics, generation history, API key management, and team settings. Live data from day one — no mock placeholders.
Authentication & Team Management
Full auth system with email/password login, email verification, password reset, and organization-scoped access control.