Roadmap
Shipped video and image workspaces, credits, and asset library — plus what's next
Backlog24
Online payment channels (WeChat / Alipay)high
More video and image modelsmedium
Credit usage insightslow
Product Listing Set: expand the category playbook library and add a mechanical pure-white main-image checkmedium
Team accounts phase 2: automatic member allowances, monthly usage export, member self-leave, multiple co-ownersmedium
Shared pool phase 2: per-member monthly / daily spending capsmedium
Viral insights phase 2: platform metrics API backfill, hand-off to the prompt wizardmedium
Analytics phase 2: in-app codebook editing, per-category facet comparison, performance correlation once platform data landsmedium
Listing image insights phase 2: fetch images by ASIN or link, per-category facets, and a bridge into the Product Listing Set agentmedium
Ops board phase 2: product / content / competitor boards, weekly creator and product ranks, buyer voice (reviews), Feishu push of new grade-A creators, Viakeo line setupmedium
Style templates and knowledge base for more categoriesmedium
Batch task queue and scheduled generationlow
Standalone pattern file export (print-ready proofing)low
More fission templates (posters / packaging)low
Broader depth-reference rollout after quality and reliability validationmedium
Extend localized render-requirements previews to more site languageslow
Project sharing and project-based My Creations viewslow
Pattern Transfer: multi-pattern x multi-colour matrix (today: one pattern per task)low
Pattern Transfer: auto-save extracted patterns into the asset library for reuselow
More audio agents (voice cloning, music, sound effects) — Text to Speech shipped as the firstlow
A batch "Generate all pattern vectors" button, mirroring the model-shot family's batch entrylow
Cover Motion Transfer with the daily production sentinelmedium
Restore the native background-source switch once the vendor implements background_sourcelow
Motion Transfer: detect the subject after upload and recommend the better modellow
In Progress0
Done74
Video Edit (modify / add / remove / re-voice / extend / splice)high
Pricing page & credit top-up tiershigh
AI video workspace (3 modes, 3 model families)high
AI image workspace (Image to Image, Text to Image)high
Credits with a signup bonushigh
Asset library and My Creationsmedium
Storyboard AI Agent (board-to-video pipeline)high
Generation timeouts with automatic full refundshigh
Fast gallery thumbnails and image zoom viewermedium
Product Listing Set agent (plan to batch render)high
Image Replica agent (batch replicate from references)high
Viral Video Remake agent (analysis, chained rendering, final cut)high
Video Upscale agent (720p–4K)high
Archive video results off vendor temp linksmedium
Bilingual user manualmedium
Garment Pattern Variations agent (smart auto-analysis, one-submit batch variants — flat-lay then model shot)high
Clear insufficient-credit feedback across every paid generation actionhigh
Protected Original/Depth remake experiment with A/B quality telemetryhigh
Localized Depth remake render-requirements previewmedium
Direct Adjust & retry across every generation workspacemedium
Per-task Kie / DeepWL choice for the Seedance fleethigh
Email-first login page with a simpler sign-in flowmedium
Compact Studio navigation with more creation spacemedium
Fictional English Logo and typography generation for every Garment Pattern Variations variantmedium
Garment Pattern Variations model list focused on HD, Nano Banana Pro, and Seedream 5 Promedium
Square 1:1 default for Garment Pattern Variations model shotsmedium
Durable free Garment Pattern Variations analysis with review-before-paymenthigh
Unified prompt-aware regeneration for Garment Pattern Variations flat-lays and model shotsmedium
Asset folder management (live)high
Browser notifications for new AI Studio releasesmedium
MiniMax H3 video generation with multi-provider routinghigh
Future fission templates reuse the smart entry (pluggable unified upload + auto-analysis)low
Seedance 2.5 video generation across the video workspace, storyboard and viral remake agentsmedium
AI Agent sidebar and dashboard home reorganized into a two-level video/image/audio navmedium
Print-ready pattern vector files (4K, 300dpi, transparent PNG) for Garment Pattern Variations and Pattern Transfermedium
Multi-select batch generation for result cards in Garment Pattern Variations and Pattern Transfermedium
Seedance 2.5 joined the Kie/DeepWL provider selector (DeepWL route live)medium
Per-user business-module visibility (admin-configurable, three enforcement layers)medium
Text to Speech audio agent: curated voices, pause/emotion markup, AI polish, per-100-char pricing, My Creations + billing wiredhigh
Voice Clone agent (personal voices wired into Text to Speech)high
Kling 3.0 and 3.0 Turbo in the video workspace (with element references)medium
Motion Transfer agent: move a reference video's motion onto a character photo, optionally matching the video's scenehigh
Wan 2.2 Animate in Motion Transfer: non-human subjects, 120-second sources, the video’s own scene nativelyhigh
HappyHorse 1.1 in the video workspace: all three modes, nine selectable aspect ratiosmedium
X Deal as a third Seedance route + platform-scheduled providers (ops-switchable, same prices)medium
Seedance prompt wizard (five-step intent capture + free media analysis + official-dialect output)high
Reference to Video: shot picker + one-click wizard analysishigh
Prompt wizard: draft health check + one-click fixhigh
Video feedback loop: 👍/👎 + issue tags, ops slices and 👎 replayhigh
Reference to Video: product-display shots anchor the first frame + shot types + single-flight detectionhigh
Reference to Video: product-swap mode (same-category demos edited object by object)high
Product Listing Set: Amazon A+ Content sets (official module specs, module copy, module-size downloads) plus plan-quality upgradeshigh
Accounts: passwordless email-code sign-up/sign-in, confirm password, auto hand-off after verificationhigh
Team accounts: owner invites members by email, allocates / reclaims credits, reviews member usage (creations isolated)high
Shared team credit pool: members spend the master’s credits and see the team balance; the master reviews per-member usage; one mode per teamhigh
Ops agent · viral video insights: break benchmark videos into 31 dimensions across 7 modules, confirm and keep themhigh
Ops agent · Analytics: a 12-field codebook plus a cross-sample analytics page (filters / library baseline / CSV export)high
Ops agent · Library scoped per team: a master account and its members share one library, and analytics cover the team onlymedium
Ops agent · Listing image insights and image analytics: batch-upload Amazon listing and A+ images by folder, 6 set-level + 7 per-image dimensions, kept and coded automatically, with analyticshigh
GPT Image 2.5 integration: Flare / Sunburst lines for both modes at GPT Image 2 pricing, with native transparent output; listing / replica gain ratio × resolution validationmedium
Ops-agent tags: breakdowns carry a required tag (e.g. "Top 100 apparel", "trending now"), lists and analytics filter by tag, older records tagged in bulkmedium
Batch production · batch video remake: one viral reference video + character / scene photos + a batch of same-style SKU images, one click renders a matching video per SKU (master prompt + per-SKU rewrite, zero interaction after launch)high
Studio home data dashboard: tool cards removed; every sign-in lands on balance and burn forecast, 7-day KPIs, running tasks, 7 / 30 / 90-day output and spend trends, where credits go, model usage, failure reasons with hints, team spend and a 30-day activity heatmap (dark / light, phone-friendly)high
Listing image insights without a per-set cap: analysed 10 images at a time, each batch saved and shown as it finishes, set-level conclusions compiled at the end, failures resume from where they stoppedhigh
Nano Banana 2 (Google Gemini 3.1 Flash Image) added to every image entry point: 5 / 8 / 12 credits at 1K / 2K / 4K, up to 14 reference images, new ultra-wide / ultra-tall 4:1 / 1:4 / 8:1 / 1:8 ratiosmedium
Batch video remake gains a Voice-over mode: the AI writes the script from the reference pacing and your product info, each SKU only swaps the product wording, with selectable VO language and subtitlesmedium
Batch video remake per-SKU product facts: each SKU’s style and selling points are recognized from its own product images (your own fields win, audience stays batch-wide), and its voice-over is rewritten from its own facts on the master’s timelinemedium
Ops board (ops agent): category lines (Vecardi / Viakeo) + the creator board — creators found daily from the product pool, graded A / B / C with category GMV and deltas, GMV per 1k followers, sales mode and e-mail, plus due-diligence metrics, 90-day trend, outreach records and CSV exportmedium
AI prompt writer in the image workspace: one line of intent plus reference images → a professional English prompt written in the selected model's official style (GPT Image / Nano Banana / Seedream), editable and applied to the prompt box with one clickmedium
Video workspace: one-click Save to library for finished videos (free, idempotent, folder picker) so a good result can be reused as a reference video in Reference to Video, Video Remake, Motion Transfer and Batch Video Remakemedium
Smart edit version history: edits no longer overwrite — the original and every edit are kept as versions you can zoom, download or save to the library, and Set as current switches back for free; covers the image workspace, storyboard boards (incl. paid re-renders) and the four image agents' result cardsmedium
Asset library Download selected: one click downloads the checked assets — a single asset as its original file, several packed into one zip of originals (no trimming or transcoding, upload names kept, duplicates numbered), failures named one by one, 500 MB per archivemedium
Color lock for Image to Image / Smart edit: corrects the result's colour cast against the reference and restores untouched areas from the reference's own pixels — deterministic, no AI, no credits, saved as a switchable version; an Auto switch runs it for results that land in this sessionmedium
Batch Production · Batch Image Swap: one finished set of shots × many SKUs in one click, each cell swapping only the product; mechanical prompts (no plan fee), per-cell charging, automatic colour lock per cell, single-cell re-render and a per-SKU zip of the whole batchmedium