Changelog

Latest product updates

Latest product updates

This week brings Google Drive shared drive support for seamless team file workflows, a major reliability pass on FAUNA chat, and dynamic generation cost display so you always know what you're spending before you generate. Plus: progress bars for uploads, canvas interaction polish, and dozens of fixes across techniques, billing, and multiplayer.



  📦 New

  • Google Drive Shared Drives — import from and export to your team's shared drives, not just personal storage. Browse and select files from any shared drive your Google account can access, making FLORA a natural part of team file workflows without manual downloads or re-uploads.

  • Dynamic Generation Costs — see accurate credit costs before you generate, updated live as you change models and parameters. No more guessing — the cost display reflects the exact model, resolution, and settings you've selected so you can make informed decisions before spending credits.

  • Upload & Download Progress Bars — precise byte-level progress replaces generic toast notifications. Track exactly how far along your file transfers are with real progress indicators instead of ambiguous spinners.

  • Batch Carousel Extract — pull individual results from batch generation carousels directly onto the canvas. Cherry-pick the outputs you like from a batch run and place them as standalone nodes for further editing or connection.

  • Community Library Emails — new and updated email notifications for technique submissions and community interactions. Stay informed when your published techniques get traction or when submissions you care about are reviewed.

  • Layer Editor: Custom Typography — The layer editor now supports custom font selection, letter spacing, and line tracking. Useful for precise text styling in branded or templated compositions.



  🧠 Improved

  • Canvas bulk delete confirmation prompt now only appears for 5+ nodes, reducing friction for small multi-selections

  • Model picker correctly filters by mention modality, so you only see compatible models when using @-mentions

  • Action node previews render at 3x resolution for sharper, clearer input thumbnails

  • Overage badge now uses warning-level styling so it's immediately visible when your workspace is in overage

  • Canvas panning works from within the marquee selection rectangle for more intuitive navigation

  • Usage breakdown now shows a clearer modality split so you can see exactly where credits go across image, video, and text



  🛠 Fixes

  • FAUNA now correctly resolves @-mentions and preserves referenced assets when triggering generations — no more broken references

  • FAUNA respects context chips you explicitly removed, so excluded context stays excluded

  • Nodes created by FAUNA properly display edges and connections

  • Technique edits reliably publish the correct graph version

  • Batch output naming no longer conflicts with auto-suffix logic

  • Right-clicking the marquee selection box no longer deselects your nodes

  • Long ProRes video exports now complete successfully with an extended timeout

  • Generations with inaccessible input images fail fast with a clear error instead of hanging indefinitely

  • Layer editor composite URLs work on view-only share links

  • Hidden projects no longer count against your active project limit

  • Project limit errors now show a friendly message instead of a raw system error

Techniques just got a major power-up: batch nodes now work inside technique graphs, so you can fan out collections, preview batch outputs, and detach techniques with all your variants intact. We also shipped a complete payment recovery flow, added Google Drive as an import source, and landed dozens of reliability fixes across generation, billing, and the canvas.

New

  • Batch nodes in techniques — build techniques that process entire collections at once. Create, preview, and run batch fanout slots directly inside technique graphs. Collection outputs now surface in the technique builder, app mode, and at runtime.

  • Fauna x Layer Editor: split to layers & fill layer to frame — Fauna now can use two new layer-editor commands for compositing workflows and quick layout adjustments. Fauna can now change aspect ratios, add text layers, resize layers, and more.

  • Google Drive import picker — pull assets straight from Drive into your canvas. Connect your Google account and browse files without leaving FLORA.

  • Gemini 3.5 Flash in Fauna — Google's latest compact model for snappier Fauna conversations.

  • Bulk-connect shortcuts — select multiple nodes and use I/V/T keyboard shortcuts to wire them up at once.

  • Usage CSV export enhancements — last-24h range option and additional columns for deeper analysis.

  • Export node actions menu — quick actions on export node inputs for faster pipeline management.

  • Krea 2 — Krea's foundation image model, built for style and aesthetic control on the canvas.

  • Opus 4.8 — Anthropic's most capable model, for even stronger text prompting.

Improved

  • Mentions & reactive canvas edges — mention upstream nodes in prompts and see live connection edges update as references resolve.

  • Cross-device model sync — your last-used model selection now syncs across devices.

  • Moderation messages — content moderation now shows specific, per-model explanations.

  • Bulk delete protection — destructive bulk node deletes now require confirmation.

  • Free plan usage indicator — the credits pill now counts up consistently, matching paid plans.

  • Personal usage limits — the credits indicator shows your individual spending limit.

  • Anthropic structured outputs — enabled the structured outputs beta with improved error handling.

  • Server-routed canvas batches — canvas batch generation runs through the server for improved reliability.

Fixes

  • Audio input previews now display correctly

  • Hex color values preserved when editing elements

  • Shape colors persist when duplicating layers

  • Spellcheck menu restored on canvas node prompts

  • Batch-imported asset toasts no longer stack incorrectly

  • Bulk edit shows shared parameters for auto-model video multi-select

  • App mode collection output previews properly ordered

  • Seedance image dimension handling fixed

This week's update brings Google Drive integration to your canvas, makes the Document Node fully production-ready, and introduces dollar-based cost tracking so you always know what you're spending. FAUNA gets smarter with layer-editor autonomy and a new default model, while batch techniques, subtitles, and a wave of canvas stability fixes round out a packed release.

📦 New

  • Export to Google Drive — sync your images, video, and audio straight to Drive without leaving the canvas. Pick a destination folder, name your files, and let resumable uploads handle the rest — including cross-batch filename deduplication so nothing overwrites. Perfect for teams that live in Google Workspace.

  • Built-in Spellcheck — real-time spell checking powered by Hunspell, running entirely in your browser. Catch typos in prompts and text nodes as you type, with no cloud round-trip and full offline support.

  • Document Node — the Document Node graduates from experimental to fully shipped with a library browser, toolbar controls, sharing, technique inputs, and loading/error handling. Spawn from thumbnails, route documents into batch workflows, and connect documents as technique inputs.

  • Add Subtitles to Video — a new "Add subtitles" utility in the video toolbar powered by Veed, making it easy to overlay text on any generated or uploaded video.

  • Dollar Cost Display — see your generation costs in real dollars. A new toggle in Preferences lets you switch the usage display from credits to USD, so budgeting is transparent.

  • Model Selection Preferences — a new "Model for new nodes" setting in Preferences lets you choose between "Last used model" (default) and a fixed default model for every new node.

  • Properties Panel for All Output Types — video, audio, and text outputs now have full properties panels, bringing them to feature parity with image outputs.

  • Audio Input Previews — see and play audio inputs directly on nodes before generating.

  • Batch Techniques — run techniques in batch mode with a dedicated run button and sibling-group output carousels, enabling high-volume variations from a single technique.

  • API Key Cap Raised to 5 — you can now have up to 5 active API keys per user, up from 1.

🧠 Improved

  • Inpaint Mask Unification — all inpainting mask methods now use a single unified pipeline, making brush, lasso, and auto-mask behave consistently.

  • Mentions Everywhere — @mentions now work in connected text-node inputs and layer editor nodes, with a smarter resolution pipeline that reacts to canvas edge changes in real-time.

  • Router Transparency — routers now behave as transparent data-flow wrappers across the canvas, improving connection rules, validity checks, and publish behavior for techniques.

  • Performance — asset list queries now stream instead of collecting all rows, and heavy re-renders during drag operations have been eliminated.

  • Better Error Messages — clearer generation model mismatch errors, accurate per-subtype moderation messages, and sanitized infrastructure errors.

  • Offline Awareness — a toast notification appears when your connection drops during multiplayer collaboration.

This weeks biggest launch: a Pricing update brings usage-based plans with included monthly credits, overage billing, and Pro/Max plans with unmetered Nano Banana 2/Pro access. Plus, five new actions, custom voices for Pro users, technique builder updates, and significant canvas performance improvements across the board.

New

  • Pricing Update — usage-based plans are here. Starter, Pro, and Max plans each include monthly usage allowances with transparent overage billing. Free users now get download and export access by default. Existing customers are automatically migrated with grandfathered rates where applicable.

  • Actions — five new prebuilt actions: QR Code Generator, Generate Text Image, Add Text to Image, Generate Shape Image, and Add Shape to Image.

  • Custom Voices for Pro — Pro plan users can now create and use custom voice profiles from ElevenLabs in audio generation workflows.

  • Technique Builder Onboarding — a new full-screen intro modal guides first-time builders through the technique creation flow.

Improved

  • Add Node Menu —    The node picker has a new layout with both list and grid views, plus tooltips that describe what each node does. Finding and adding the right node is faster and more discoverable across the full node type catalog.

  • Technique Library: Favorites — Techniques can now be starred as favorites directly from the library picker. Favorited techniques float to the top of the list, making frequently-used workflows faster to access.

  • Batch Rerun from Masonry — rerun individual cells directly from the masonry grid view without switching context.

  • Copy-Paste with History — pasting nodes now duplicates full generation history, so you keep your creative trail intact.

  • Technique Builder — fit/fill toggle for thumbnails, test-run errors visible in the workflow overlay, and test-run results on the Status tab.

  • Canvas Performance — targeted node and edge selectors, optimized hover animations, throttled text-block resize, and faster connection line rendering for smoother editing on large canvases.

  • Generation Loading — smoother progress indicator and loading state that now survives page refresh.

  • Audio Nodes — drop and replace support for audio nodes, audio playback stays alive across generation switches, and downloads preserve the correct file extension.

  • Gemini Seed Parameter — forward seed values to Nano Banana for deterministic, reproducible generations.

  • All-you-can-eat Nano Banana — Pro and Max plans now include unmetered generations on Nano Banana models with higher rate limits (225/hour).

Fixed

  • Canvas zoom speed normalized when using Ctrl+scroll

  • Image crop dimensions now floor correctly to maintain aspect ratio within provider bounds

  • Long technique names properly truncated in the node picker dropdown

  • Canvas edges now respect model input limits on body-drop

  • Inpaint mask outline restored for outline composite endpoints

  • Video model picker correctly scoped when both audio and image/video inputs are connected

  • Pasted technique output connections are now preserved

  • Orphan edges cleaned up from canvas rendering

  • Label layering fixed after ungrouping nodes

  • Outpaint source geometry corrected

  • App mode: cleaner workflow view and direct single-output downloads