juin 2026
Features
What you can do now
WHAT YOU CAN DO NOW
The Metadata Enrichment Agent fills your DAM metadata fields from AI prompts you write, based on what each asset shows.
→ Works with your field types. True/false values, numbers and predefined lists: the agent follows your metadata configuration and your prompt.
→ Your own classifier in minutes. Build a custom asset classifier in a few clicks, or set up bounded tags in under a minute.
You asked, we built: Metadata is what makes assets findable, and libraries keep growing. Now the agent fills it in as they do.
Works well with: AI Trigger Rules
What you can do now
WHAT YOU CAN DO NOW
On the post-process page, AI Trigger Rules decide whether a process runs based on what an image actually shows, not just its file properties.
→ Chain rules. Combine several rules for more complex workflows, so automated processes run only when your visual criteria are met.
You asked, we built: Not every process fits every image. Now each one runs where it’s useful.
Works well with: Metadata Enrichment Agent · Skip Remove background for transparent images
Improvements
What you can do now
WHAT YOU CAN DO NOW
The Face Matcher agent’s recognition engine has been upgraded for real-world images. It now holds up across tricky orientations, angles and face sizes.
You asked, we built: Event, sports and campaign photos rarely show faces straight on. Matching now handles them.
Works well with: Metadata Enrichment Agent
avril 2026
Features
What you can do now
WHAT YOU CAN DO NOW
The Scaleflex asset picker is now built into the MCP server. In chats with MCP-compatible assistants, upload and preview assets, edit metadata on the fly, and open any asset in Hub in one click.
You asked, we built: Working with images in a chat is easier when you can see them.
Works well with: Manage your library from an AI assistant
Improvements
What you can do now
WHAT YOU CAN DO NOW
Customers with high-volume use of advanced Visual AI features can now run on dedicated infrastructure sized to their demand, with stable throughput.
You asked, we built: Large catalogues and campaign peaks need predictable processing capacity.
Works well with: External Services Configuration
mars 2026
Features
What you can do now
WHAT YOU CAN DO NOW
Visual AI now integrates Google’s Nano Banana 2 and Nano Banana Pro models for image generation and advanced editing. Make precise changes to an image, or create a new asset from a text prompt.
→ Retail. Standardise imagery for regulatory compliance.
→ E-commerce. Stage products virtually from simple pack-shots.
→ Real estate. Enhance property photos to show each listing at its best.
You asked, we built: Teams wanted generative editing in the same place their images already live.
Works well with: Expand Background based on ratio
Improvements
What you can do now
WHAT YOU CAN DO NOW
The Scaleflex MCP server, built on the Filerobot API, now covers more of asset management.
→ Tags. Create, attach and delete them.
→ Metadata. Edit values directly.
→ Assets. Upload from a URL, and delete.
You asked, we built: Teams using AI assistants wanted to run library tasks from the chat, not only look things up.
Works well with: Pick and preview assets inside your AI chat
What you can do now
WHAT YOU CAN DO NOW
The external services Visual AI relies on are now configured to use European servers whenever possible, in line with European data protection law.
You asked, we built: Customers in Europe want their data processed in Europe. Now it is, wherever possible.
Works well with: Custom Dedicated Routing
février 2026
Features
What you can do now
WHAT YOU CAN DO NOW
Visual AI can now detect whether the five-colour Nutri-Score rating is present on product packaging.
→ Audit whole catalogues. Run it across your product images to check that every item meets mandatory labelling requirements.
You asked, we built: For retailers, labelling compliance has to be confirmed product by product. Now it can be checked automatically.
Works well with: AI Trigger Rules
What you can do now
WHAT YOU CAN DO NOW
Expand Background has a new option: set a target aspect ratio, such as 16:9, 4:3 or 1:1, and the background extends to match, instead of setting dimensions by resolution.
→ Fit every platform. Social posts, banners and wide headers all get the format they need, and the original subject stays intact.
You asked, we built: Each channel has its own format. Now one image fits them all.
Works well with: Skip Remove background for transparent images
Improvements
What you can do now
WHAT YOU CAN DO NOW
A new parameter detects transparent pixels. When an image already has transparency, Remove Background is skipped entirely.
You asked, we built: Cut-outs often go through the same pipeline as every other image. Now they don’t spend credits on work that’s already done.
Works well with: AI Trigger Rules · Template Builder - Require background-removed version