FrameRef 1.1 opens portals across your reference mess

FrameRef 1.1 tackles large reference collections with portals, voice memos, mind-map edges, labels, AI tagging, image tools and improved video navigation.
A dark design workspace shows a selected landscape-rendered frame preview beside a 3D preview panel and technical floor plan drawings. The central image features a minimalist black cabin glowing warmly beside still water, with its reflection and misty mountains in a cool blue-gray scene.

In case you don’t know the tool: FrameRef is Polygonflow’s reference-board application for Windows and macOS. It puts images, videos, PDFs and GIFs onto one infinite, zoomable canvas, where they can be arranged, grouped, annotated and transformed non-destructively. References can be brought in through drag and drop, the browser or clipboard, while tools for cropping, scaling, rotation, flipping, snapping, text notes, drawing and Always on Top turn the board into a working visual reference space rather than a particularly ambitious Downloads folder. We covered FrameRef 1.0 in detail here.

At some point on every carefully curated reference board, “carefully curated” becomes a joke. The images multiply. Videos arrive. PDFs settle in for the winter. Somewhere in the northeast quadrant is that one reference you definitely need, assuming you can remember where northeast is. FrameRef 1.1.0 has arrived for precisely this stage of reference-board civilisation.

It adds AI tools, voice memos, mind-map connections, portals and a customisable toolbar, but the more interesting theme is organisation. Several of the new tools are designed to make a large board easier to navigate and search once it has grown well beyond moodboard size.

Portals for the infinite canvas

Portals are a new FrameRef item type intended specifically for navigating large boards.

A dark design canvas shows a selected purple oval brush stroke inside a blue resizing box, with corner handles and a size label reading 844 × 844. A floating toolbar above includes an Edit label button, a star icon, a tag icon, and a lock icon, all set against a black background with a partial teal curve in the corner.

A portal links location A on the board with location B. Enter Portal Mode and the canvas remains navigable with MMB drag for panning and the scroll wheel for zooming. LMB places the first portal and then its second, linked counterpart. Once the pair exists, select either portal and press T, or use the teleport button in the context-sensitive menu. FrameRef moves the view to the linked portal. No Wheatly-Voiceover, but…

More importantly, a portal can retain a particular view rather than merely pointing towards another coordinate. Its pan and zoom can be updated using the square-with-a-dot control in the context-sensitive menu. That means a sprawling board can effectively acquire destinations: characters over here, architecture over there, lighting references somewhere beyond the horizon, and no need to spend half the afternoon zooming through the void like a underfunded space programme.

Mind maps join the board

FrameRef 1.1 also introduces Mind Map Edges. Hover over an item and a blue dot appears. Dragging from that dot creates a connection to another item, allowing references to be linked visually rather than merely positioned near each other.

A dark interface shows two reference images connected by curved blue arrows, suggesting a linking or relabeling workflow. On the left, a full-body photo of a man in a hat stands against a wooden wall; on the right, a flat lay of a white T-shirt and brown pants sits on a light surface.

It is a small distinction until a board becomes large. Spatial proximity says “these happened to end up together”. An explicit edge can say “these actually belong together”. FrameRef does not describe this as a dependency or production graph. It is a visual mind-map system.

Labels are becoming more general as well. The context-sensitive menu now contains Edit Label for all item types, giving the board another organisational layer without requiring everything to be expressed through groups and geography. And if geography fails entirely, there is search.

AI tags your pile

The first of FrameRef’s new AI functions is AI Tagging. Select one or several images on the canvas and run Tag Selected Images. FrameRef automatically generates several search tags and attaches them to the selected images.

Those tags feed FrameRef’s existing search, so the practical benefit is retrieval rather than image generation. For anyone who collects references with the rigorous metadata discipline of “drag it onto the board and deal with it later”, this may be one of the more useful AI additions.

A dark-themed image tagging interface shows a grid of photo thumbnails, with an ‘Edit tags’ pop-up centered over the canvas. The surrounding thumbnails include a standing man indoors, Santa Claus in a snowy scene, a blue sky, and a waterfront town, while a smaller AI Tagging panel sits to the right, all lit by muted charcoal panels and soft blue-gray accents.

Image Edit takes between one and five selected images. Hit AI Edit, provide a text prompt, and the tool edits the existing imagery while attempting to preserve its composition. Image Gen is accessed through FrameRef search and generates an image from a description.

Image Try-On is more specialised. It is designed to virtually fit one or more garments onto a photograph of a person. The model requires a prompt plus a person image and a garment image. Images can be dragged between the Person and Garment fields inside the editor. Holding Ctrl also allows images to be dragged directly from the FrameRef board onto the AI editor window.

Image Upscale operates on one selected image and increases it to a chosen target resolution. Polygonflow specifies a maximum output of 128 MP. Video Animate takes a still image and a reference video. The image supplies what is being animated and the video supplies the motion reference. The tool is launched through FrameRef search.

AI gets its own wallet

These operations are not unlimited additions to the FrameRef licence. Polygonflow says the AI models cost money to run, so FrameRef 1.1 introduces a separate credit system. The new AI menu links to Polygonflow’s AI platform, where information about the available models can be viewed.

A dark-themed financial dashboard titled Add Funds shows a balance summary, purchase credit buttons, and a transaction history table beneath. The interface uses charcoal panels, muted gray text, and red status tags, with a sparse layout framed by slim top navigation and a black background.

The Dashboard provides account creation, funding and transaction information. Credits are purchased as one-time packages rather than through a recurring AI subscription described in these release notes. After payment is completed and the funds have appeared in the account, those credits can be consumed by FrameRef’s AI functions. Or you can just not sue the AI, and save the money.

Voice notes for when typing is too civilised

Voice Memo is another new tool available through FrameRef search. Recording happens directly inside FrameRef. The Voice Memo window can start and stop recording, play it back or throw it away. Input volume is adjustable and the recording device can be selected, which matters on a workstation containing six microphones, three virtual audio devices and one webcam that Windows remains mysteriously fond of.

Recorded audio can also be trimmed using yellow brackets. For reference work, this is potentially more useful than it first appears. A picture tells you what you collected. A ten-second memo can tell Future You why you collected it, what part matters and why the apparently unrelated image next to it is not, in fact, evidence that Past You had finally lost the plot. Or, if you get the File with the Art Director standing behind you, and they are describing what they want, that might be a better way to record than trying to understand their input during typing.

More precise video navigation

Video navigation gets several changes. The commands for jumping forwards and backwards now use larger increments. For video, the jump is five seconds. For sequence-style items such as PDFs and GIFs, it is ten frames. Videos also gain Go to Time. Right-click a video, select the command and enter the exact time you want to reach rather than scrubbing until the correct moment eventually wanders past.

From moodboard to reference library

FrameRef 1.0 established the basic idea: one infinite, zoomable board that can hold images, videos, PDFs and GIFs together. References can arrive through drag and drop, the browser or clipboard, and items can be scaled, rotated, cropped and flipped non-destructively. There are also grouping, snapping, grid alignment, text notes, colours, drawn annotations and an Always on Top mode for keeping the board visible beside another application.

Version 1.1 starts dealing with what happens after you actually use such an infinite board for a while. Portals provide destinations. Mind Map Edges provide relationships. Generic labels and AI tags provide another route to finding things. Voice memos preserve the thought that made a reference useful in the first place. The AI editing and generation tools can then modify or create material without leaving FrameRef.

Individually, these are quite different additions. Together, they push FrameRef away from being merely a large surface on which to throw pictures and towards something closer to a navigable reference library. Which is probably sensible. Give creative people an infinite canvas and they will eventually discover whether “infinite” was a promise or a threat.

Availability

FrameRef 1.1.0 can be downloaded and installed through Check for Updates in the Help menu. Polygonflow also provides separate Windows and macOS downloads.

Source: Polygonflow FrameRef 1.1.0 release notes