What's new

OpenAI's flagship image model lands on MoodNode via fal.ai — state-of-the-art photorealism, pixel-perfect text rendering, brand-consistent product photography, available at 1K / 2K / 4K. Pricing now scales correctly with resolution across all multi-resolution models, fixing a long-standing under-charge on Banana Pro and Banana 2 4K outputs.

• ✓GPT Image 2 — text-to-image with photorealistic output (1K / 2K / 4K)
• ✓GPT Image 2 Edit — reference-based editing with mask support for inpainting/outpainting
• ✓Resolution-aware credit pricing — credit cost now scales with the resolution toggle
• ✓Banana Pro 4K and Banana 2 (2K / 4K) now charged at the correct provider rate
• ✓New `audit-pricing` script + verified-source manifest catches future provider price drift

Big behind-the-scenes upgrade to how the canvas handles your work. Every destructive action — delete, duplicate, multi-select, keyboard shortcuts — is now fully undoable, and opening the same project on two devices won't silently overwrite your changes anymore. Plus a smoother dashboard, clearer offline state, and faster project recovery.

• ✓Undo now covers every action — delete, alt-drag duplicate, Cmd+A, Cmd+D, bulk select
• ✓Concurrent-edit protection — edit on two devices and you'll see a conflict banner instead of losing work
• ✓Dashboard skeleton loader + clearer error messages when a project fails to load
• ✓Global offline indicator so you know when the network drops
• ✓ZIP template import — full round-trip with exported templates
• ✓Projects no longer auto-regenerate on open — only when you ask
• ✓Tiered snapshots — up to 30 recovery points per project, automatically

Three new engines land this week. Kling Image V3 and V3 Edit bring high-fidelity image generation and editing. Seedance 2.0 and Wan 2.7 join the video lineup via OpenRouter. Credit previews now show the exact cost for every supported duration, and Wan models auto-adjust unsupported aspect ratios instead of failing.

• ✓Kling Image V3 — high-fidelity image generation
• ✓Kling Image V3 Edit — precise instructed edits on existing images
• ✓Seedance 2.0 via OpenRouter — new fast video model
• ✓Wan 2.7 via OpenRouter — upgraded from Wan 2.5
• ✓Precise credit preview for every model's supported duration range
• ✓Wan models now auto-adjust unsupported aspect ratios instead of failing

Bring your own Google Gemini API key and generate Veo 3.1 video directly — you pay Google's API rates with zero MoodNode markup. Veo 3.1 Lite also added as a Vertex AI binding for the budget tier. Perfect for heavy video users who want the most cost-efficient path to Google's latest video model.

• ✓BYOK Veo 3.1 via Google Gemini API — pay Google's rates, zero markup
• ✓Veo 3.1 Lite on Vertex AI — budget tier with the same BYOK path
• ✓Secure key storage — encrypted at rest, never leaves your browser unencrypted
• ✓Automatic routing — BYOK when you have a key, platform credits otherwise

The landing site is significantly faster now — all heavy assets (videos, prompts data) are served from Cloudflare R2 with aggressive compression. Prompt library also got a Still/Motion filter and video previews on hover so you can see motion prompts come alive before opening them.

• ✓Landing pages and media now served from Cloudflare R2 — faster everywhere
• ✓Feature videos compressed ~94% with no visible quality loss
• ✓Prompt library: Still vs Motion filter
• ✓Hover-to-play video previews on motion prompts
• ✓Google Translate no longer crashes nodes

Rebuilt how MoodNode tracks every AI model behind the scenes — one source of truth means every new model ships with correct pricing, capabilities, and provider routing automatically. On the surface: audio on/off pricing now shows on the landing page, provider fallbacks notify you via toast, and gallery navigation gracefully falls back to the last valid image.

• ✓Single source of truth for every AI model — faster, safer model updates
• ✓Audio on/off pricing visible on landing pricing table
• ✓Toast notification when a provider fallback activates (no silent swaps)
• ✓Gallery navigation keeps the last valid image if one fails to load
• ✓Ideogram V2 now supports character reference images

Add your OpenRouter API key and unlock hundreds of LLMs plus image generation through a single provider — Claude, GPT, Gemini, Llama, and more, all billed through your OpenRouter account with zero MoodNode markup. One key, the entire ecosystem.

• ✓BYOK OpenRouter — one key, access to every model OpenRouter supports
• ✓LLM nodes route to OpenRouter when your key is set
• ✓Image generation models also supported via OpenRouter
• ✓Zero MoodNode markup — pay only what OpenRouter charges
• ✓BYOK mode now persists across sessions — no more re-enabling on refresh

Tightened up the rough edges from this week's launches. Director's Eye motion videos now play smoothly on hover, the wordmark logo renders crisp instead of rasterized, and the in-app MoodLibrary picked up a Director's Eye tab so logged-in users find the techniques without leaving the app.

• ✓Director's Eye motion video playback fixed in nav menu and gallery
• ✓Logo SVG renders sharp at every zoom level (no more rasterization)
• ✓Director's Eye now appears as a third tab inside the app's MoodLibrary
• ✓Prompt detail pages — title and actions moved below the canvas preview
• ✓Bigger thumbnails on in-app prompt cards

A curated cinematic prompt library covering 83 still techniques and 18 motion camera moves. Each technique includes a real video example, when-to-use guidance, and pro tips — designed to teach the language of directors of photography while you generate.

• ✓101 cinema technique prompts with real generated examples
• ✓18 motion prompts with looping video demos for camera moves
• ✓Immersive gallery view with hover-to-play motion previews
• ✓Per-technique detail pages with When to Use and Pro Tips
• ✓One-click Open in Canvas deep link — opens the right node type per prompt

The MoodLibrary is now a standalone discovery surface on the landing site — 324 curated AI prompts with thumbnails, categories, and full-text search. Click any prompt and Open in Canvas drops you straight into the app with the prompt and an image node already wired up.

• ✓Public prompt library at moodnode.ai/tools/prompt-library
• ✓324+ curated prompts across 10 categories with search
• ✓Per-prompt detail pages with copy, share, and Open in Canvas
• ✓Open in Canvas deep link auto-creates the project and lands you in the editor
• ✓Category landing pages with SEO metadata for each style

Full-site SEO/GEO optimization, dedicated FAQ and comparison pages, and a hero LCP that dropped from 4.2 seconds to roughly 0.1. Smaller fixes across the editor too — sidebar group nodes, compare node sizing, edge alignment, HEIC drag-and-drop.

• ✓335+ indexed pages with structured data and OG metadata
• ✓Dedicated FAQ page (23 questions, 6 categories)
• ✓Comparison pages for ComfyUI, Midjourney, and Leonardo
• ✓Hero h1 server-rendered — LCP 4.2s → ~0.1s
• ✓Disposable email signup gate (client + server-side)

Six new video models added. Google's budget Veo 3.1 Lite, PixVerse v6 with style presets, Kling V3 Pro & Standard with audio and voice control, and Wan 2.5 replacing older versions. Updated pricing across the board.

• ✓Veo 3.1 Lite — budget Google video, 720p/1080p
• ✓PixVerse v6 — styles (anime, cyberpunk, clay), 1080p
• ✓Kling V3 Pro & Standard — audio, voice control, 1080p
• ✓Wan 2.5 replaces Wan 2.2/2.6 — auto-migration for existing projects
• ✓Smart duration clamping for all Veo models

Full email composer in admin dashboard with template picker, live HTML preview, and auto credit grants. Bug Hunter template rewards users who report bugs with credits automatically.

• ✓Mailing tab — template list + live preview side by side
• ✓Editable templates — pre-filled subject & body, customize before sending
• ✓Bug Hunter template — auto-grants 100 credits on send
• ✓Markdown formatting in emails — **bold**, *italic*, __underline__
• ✓Mail assets organized in /mail/ directory with 301 redirects

Video nodes now match image nodes with proper hover overlays showing model, prompt, and gallery navigation. Multiple crash fixes and canvas improvements.

• ✓Video overlay — model name, prompt, delete button (matches image node)
• ✓Gallery navigation — ChevronLeft/Right arrows, centered layout
• ✓Shift + Click multi-select (alongside Cmd)
• ✓Fixed infinite loop when connecting Last Frame output
• ✓Hidden input handles no longer push visible handles down
• ✓Share modal — text selection no longer closes the dialog

Share your workflows with a public link. Free users get 90-day links (max 5), subscribers get unlimited. Plus, daily free agentic workflow builds and improved canvas controls.

• ✓Share links — public URL with expiry management
• ✓Expired & not-found pages with custom designs
• ✓Agentic Workflow — 5 free builds/day, unlimited for subscribers
• ✓Canvas zoom controls & empty state hints
• ✓Share tracking — view counts, referral attribution
• ✓Admin share management panel

Display all your generated images in one place. Connect any image node to Gallery and see results in a masonry grid. Plus, the sidebar and Quick Add menus now show the exact same 40+ nodes — nothing hidden.

• ✓Gallery Node — masonry grid, Download All, Save to Collection
• ✓Sidebar and Quick Add unified — same nodes, same categories
• ✓Keyboard shortcuts in dashboard help menu
• ✓Performance boost for large workflows

Generate images and videos with Runway's latest Gen-4 model. Connect a reference image and prompt to create cinematic video clips with advanced motion control.

• ✓Image and video generation
• ✓Reference image support
• ✓Multiple aspect ratios and durations

No more waitlist — anyone can sign up and start creating. Flexible pricing from free to Business, with 150 starter credits for every new account.

• ✓Free plan with BYOK (Bring Your Own Key)
• ✓5 plans: Free, Lite ($4), Starter ($9), Pro ($24), Business ($49)
• ✓One-time credit packages and custom slider
• ✓150 free starter credits for everyone

Seedream 5.0, Recraft V4, Qwen Image, Grok Imagine, and 8 more. MoodNode now supports 50+ AI models for image, video, text, and audio — all in one workspace.

• ✓Seedream 5.0 by ByteDance
• ✓Recraft V4 for design assets
• ✓Qwen Image 2 Pro
• ✓Grok Imagine by xAI
• ✓Updated video pricing tiers

Browse generated videos in a fullscreen gallery with hover preview. The model selector now shows pricing and capabilities so you know the cost before generating.

• ✓Fullscreen video gallery with playback controls
• ✓Hover preview on video thumbnails
• ✓Model pricing displayed before generation
• ✓Favorite models for quick selection
• ✓Microsoft Clarity for session analytics

Reference any node with @nodename in your prompt — images are automatically pulled as references. Save and reuse styles with /stylename syntax.

• ✓@nodename — pull images from any node as reference
• ✓/stylename — apply saved styles to prompts
• ✓Style autocomplete with search
• ✓Custom style save from prompt library

Five new AI-powered edit nodes. Paint a mask to edit areas, control pose and depth, expand boundaries, extract objects, or convert to SVG.

• ✓Inpainting — paint mask, describe replacement
• ✓ControlNet — pose, depth, canny edge, line art, normal map
• ✓Expand Image — AI outpainting
• ✓Segmentation — SAM-powered object extraction
• ✓Vectorize — SVG conversion (moved from v0.4)

Share workflows with a public link — viewers can clone and remix. Dual credit system: plan credits refresh monthly, purchased credits never expire. Full security audit passed.

• ✓Public share links with expiry control
• ✓One-click clone to remix shared workflows
• ✓Plan credits + bonus credits (never expire)
• ✓Security audit — all critical findings resolved

Complete landing page redesign with new hero section, pricing grid, FAQ, testimonials, and blog. Polar.sh payment integration for subscriptions and credit purchases.

• ✓New landing page with interactive hero demo
• ✓Pricing page with plan comparison
• ✓Polar.sh payment integration
• ✓Blog with SEO and structured data

Share your workflows with anyone via a public link. Shared canvases show your nodes and connections as read-only, with the option to clone and remix.

• ✓Public share links with expiry
• ✓Clone shared workflows with one click
• ✓OG preview images for social sharing
• ✓WebP image optimization for shared content

Organize reference images into collections with drag & drop. Image nodes now accept multiple reference images — connect as many as the model supports.

• ✓Collection node — drag & drop, multi-select, per-model ref limits
• ✓Dynamic reference inputs on image and video nodes
• ✓Pinterest-style collection naming

Write prompts directly inside image and video nodes — no need for a separate Text node. Plus, redeem promo codes for free credits during alpha.

• ✓Inline prompt — write directly in generation nodes
• ✓Promo code redemption for alpha credits

Plan visual stories with the Storyboard mode in LLM Enhance. Create sprite sheets from any image — split frames, animate preview, export as GIF or atlas for game development.

• ✓Storyboard mode — visual scene planning with LLM
• ✓Sprite Sheet — frame splitting, animation preview, GIF export
• ✓Promo code system

Upscale videos with AI-powered super resolution. Build complex camera prompts with the new Camera Preset node — combine style, lens, movement, and lighting.

• ✓Video Upscale — 2x and 4x AI super resolution
• ✓Camera Preset v2 — style, lens, movement, lighting builder
• ✓Project snapshots for version safety

Invite-only alpha launch gate, cookie consent, email verification, and legal pages. Sentry error tracking and What's New changelog panel.

• ✓Invite-only alpha launch gate
• ✓Cookie consent (GDPR/KVKK)
• ✓Terms of Service, Privacy Policy
• ✓Sentry error tracking

Full cloud save with automatic sync. Projects are saved to Supabase Storage with quota enforcement and incremental dirty-node tracking.

• ✓Cloud save with auto-sync
• ✓Supabase Storage with quota enforcement
• ✓LRU cache eviction for blob management

Convert raster images to SVG with the Vectorize node. New animated neural pulse edges visualize data flow between connected nodes.

• ✓Vectorize — raster to SVG conversion
• ✓Neural pulse edge animations

Three new AI Tools nodes. Sync audio to faces with LipSync, transfer motion between characters, and swap faces in images.

• ✓LipSync — audio + face synchronization
• ✓Motion Control — transfer motion from reference video
• ✓Face Swap — swap faces in generated images

Canvas-based storyboard for scene planning, image enhancement tools, and UX improvements across the workspace.

• ✓Storyboard canvas for scene planning
• ✓Image enhancement tools
• ✓Character sheet and body sheet prompts

MoodNode starts as a node-based visual AI workspace. Connect AI models with drag-and-drop to create images, videos, and text — no coding required.

• ✓50+ AI models (FLUX, Kling, Stable Diffusion, GPT-4, Claude)
• ✓Visual node editor with real-time connections
• ✓Crop, Adjust, Enhance, Export nodes
• ✓Text Array for batch prompts
• ✓BYOK — use your own API keys at no cost