Designer

The Designer (/media/designer) is Postmill's manual canvas editor, built on Konva / react-konva — it is not Polotno. It's the one Platform-section tool with no header in the nav rail, and it's the target every other studio's Edit in Designer button hands off to.

Two modes

Mode	Use	Output
Static canvas	Images, text, shapes on a single frame.	PNG / JPEG / WebP / transparent PNG / PDF / SVG.
Video timeline	Multi-track video composition.	MP4 / WebM / GIF / animated WebP.

Animated GIF and animated WebP export only exist in video mode — static mode flattens the canvas to a single frame, so those two formats aren't offered there.

A new session always opens a required Start a design dialog (no silent blank-canvas default): open something from Recent Designs or Templates, or start fresh. New Photo lets you pick one or more channel-preset sizes (or a Custom size…) — each becomes a linked output tab. New Video takes one format to start; more video formats can be added later as output tabs. Untitled designs are saved automatically on your first edit — an untouched blank canvas is never saved.

Formats, pages and versions

• Output tabs — a design can carry several formats (sizes) at once. Add format adds another size; linked elements stay in sync across formats unless you choose Edit Format Only.
• Pages — image outputs can hold multiple pages (up to 20 per output), managed in the Pages panel (add, duplicate, rename, delete). PDF export combines an output's pages into one file.
• Versions — File ▸ Save Version… snapshots the design; the Versions panel lists manual and automatic snapshots and can restore or delete them.
• Templates — Save as Template stores the current design (image or video) as a template; Browse Templates… filters templates by category and organises them into folders. Applying a video template switches the design to video mode (and vice versa).

Video timeline

The timeline supports eight track types: video, image, text, shape, paint, caption, audio, and sticker, up to 40 tracks, with a composition limit of 180 seconds. Audio waveforms and video filmstrips are drawn client-side so you can see content while trimming. Rendering runs server-side via headless Chromium + FFmpeg.

• Tracks — add tracks with + Track, or drag a clip below the last track to create a new one. Tracks can be renamed, muted, locked, and hidden. A dropped clip snaps into the nearest free gap on the target track; if there is no room it is rejected rather than overlapping (hold Alt to keep an intentional overlap).
• Editing — split at the playhead (S), trim, duplicate, freeze frame (Ctrl+F), ripple delete (Shift+Delete), multi-select (Shift-click or marquee) with copy/paste, snapping, timeline markers, zoom, and loop playback.
• Transitions — cut, fade, dissolve, slide, wipe, zoom, and blur between clips.
• Clip inspector — speed and speed ramps, reverse, volume (keyframable), fade in/out, auto-duck under voice, Looks presets, filters, blend modes, motion blur, Ken Burns presets, and per-property keyframes with a graph editor. Animations presets apply to text, shape, image, and sticker clips; text clips also get a Typewriter reveal. Audio clips can detect beats.
• Captions — CC auto-generates captions from every audio-bearing clip (needs a speech-to-text provider such as Deepgram), imports .srt/.vtt files, and exports the caption tracks as .srt.
• AI generation — the timeline can generate an AI Video, Music, a Voiceover (generated, or recorded with your microphone), an Avatar video, or a Slideshow, and can transform, upscale, remove the background of, or regenerate a segment of a selected clip. Each generate dialog has a Provider: picker: Default uses your workspace's media default, or pick a configured provider to use its compact studio inline.

Canvas tools

Beyond layers, shapes, text, and pen/paint tools, the Tools menu carries the AI image actions — Generate Image…, Inpaint…, Remove Background, and Expand Image (AI)…, which extends an image beyond its borders (zoom out 1.5×/2×, make square, or extend left, right, up, or down) and requires Replicate. The color picker supports alpha; fills include patterns (with a noise preset) and output backgrounds can carry grain or paper textures. The font picker includes Upload font for your own .ttf, .otf, or .woff2 files (WOFF2 fonts are not used in server-rendered exports).

Opening the Designer

The Designer accepts a few different entry points, all via the /media/designer route:

Entry point	How it arrives	What loads
Single asset	?url=&type=&w=&h= query params (from a studio's Edit in Designer, or a stock browser's Open in Designer).	An image drops onto the static canvas as a thumbnail; audio/video route to the timeline handoff below.
Bulk handoff	Files → select multiple → Open all in Designer stashes the selection in sessionStorage and navigates to ?bulk=1.	All selected assets load onto the static canvas at once.
Timeline handoff	Any studio's render-queue Edit in Designer for audio/video stashes { type, url, fileId } in sessionStorage and opens ?timeline=1.	The asset loads directly onto the video timeline.
Caption handoff	Deepgram's Edit in Designer button (video sources only) stashes { url, fileId, width, height, words } and opens ?captions=1.	The video loads onto the timeline and a caption track is built from the word timings — no re-transcription.

The caption handoff is the only path (together with the general timeline handoff) that puts a real video clip on the timeline from a URL — a plain ?url=&type=video single-asset open only drops a static thumbnail, not a timeline clip.

Menu bar and shortcuts

The full action set — file, edit, insert, format, and export commands — lives in one registry that drives both the top menu bar and the ⌘K command palette, so every action is reachable both ways. Keyboard shortcuts are scoped to canvas focus. On mobile, the menu bar collapses to four grouped menus plus a ☰ overflow.

Exporting

The Export dialog offers different formats per mode:

• Static canvas: PNG, JPEG, transparent PNG, WebP, PDF, SVG.
• Video timeline: MP4, WebM, GIF, animated WebP.

Video exports also take a resolution (720p / 1080p / 4K shorthands or a free scale), frame rate, Platform presets (TikTok, Reels, Shorts, YouTube), optional Normalize loudness (-14 LUFS / -1.5 dBTP, the social-platform target), and a poster/thumbnail. Video renders run asynchronously; progress shows in the dialog.

GIF and animated WebP are deliberately excluded from the static-canvas format list — a single-frame Konva snapshot can't produce real animation, so those two only exist as video-mode exports.

Exports can save back into /files (with folder selection) or post directly into the composer. After a static-canvas export, the done step also lists each exported file as a download link so you can save a copy to your computer.

---

Verified against v1.3.0