WorkBuddy Doesn't Edit for You, It Organizes Your Asset Inventory

Conclusion first: The more browser-based video editors like UNICUT resemble desktop software, the more WorkBuddy should step forward to manage assets and delivery status. I stumbled upon UNICUT today—an open-source Web video editor with multi-track timelines, keyframes, subtitles, background removal, and even AI Agents that can operate the editor. It's suitable for editing wedding short films and e-commerce product videos. But what I really want to discuss is: before it edits, have I organized a messy pile of files into a deliverable ledger?

I've used WorkBuddy for about a month. WorkBuddy is an AI office assistant that can read documents, organize spreadsheets, handle email attachments, and break down tasks. It can't drag that second of fade-in on the timeline for me, but it can compress photos from WeChat, client emails, footage shot by photographers, and assistant revision comments into a single table. I call this table a "ledger," essentially a master table of assets and delivery status. Without it, any editing tool becomes a firefighting scene.

My setup is quite basic. Open WorkBuddy, create a new workspace named "Video Delivery Ledger." A workspace is like a project room, holding files, tasks, tables, and member permissions. Click import and throw in the asset folder, Excel list, and email attachments together. Don't be greedy here; don't say "organize everything" right away. First create a template with fixed fields including Original Filename, Source, Format, Duration, Visual Issues, Client Version, Acceptance Status, Person in Charge, and Deadline. Fields are the columns in the table. Set the Format column as an enum, allowing only mp4, mov, jpg, png, pdf, psd. An enum limits options, preventing assistants from filling in things like "WeChat video," "that HD one," or "the dusk shot."

I also fix my prompts: classify by filename and format, keep only one copy of duplicate assets and mark the path, do not guess unrecognized items but put them in "To Be Confirmed"; every record must retain the original filename, do not rename. After running, WorkBuddy generates a summary table. I hit a pitfall the first time: with mixed formats, it would mix PDF headers with Excel column names, resulting in a table that looked like misaligned Sudoku. Later, after every import, I added a field mapping step, manually matching recognized columns to template columns, e.g., grouping "Filename," "Original Name," and "Asset Name" into the same column. This step takes about two minutes but saves thirty minutes of table alignment later.

I also divide permissions into three categories. Open WorkBuddy's workspace settings, go to member permissions. Assistants can edit asset lists and "To Be Confirmed" items; editors can only modify editing progress and issue notes; clients have read-only access to a delivery status view. A view is a table filtered by conditions. Clients only see Passed, Pending Review, and Re-edit; they don't see original paths or assistant notes. Wedding clients fear a comment like "can you fix this face" messing up versions; e-commerce clients fear "this gradient looks like dusk color temperature" being interpreted as adding filters arbitrarily. Letting them see only status reduces a lot of bickering.

Daily operations are fixed. Five minutes before wrapping up each day, let WorkBuddy sync shared folders and email attachments to generate today's new assets. Export the ledger every Friday and save it to the project archive directory. Change the status to Closed for projects that have passed acceptance to prevent them from appearing on the homepage. During peak seasons, we might receive over two hundred images and dozens of videos a day; without closed statuses, the task list grows thicker, looking like a film negative box spilled on the floor.

Back to UNICUT. It puts multi-track timelines, asset panels, and export flows into the browser. Multi-track timelines allow layering visuals, music, and subtitles. This tool is suitable for execution, but I don't hand over delivery judgments to it. WorkBuddy first clarifies which images can move, which videos to edit first, which version the client nodded at, and which needs re-editing. When handing off to editors, WorkBuddy outputs a task card containing asset name, duration, in-point, out-point, cover suggestions, and revision notes. Editors then enter UNICUT and drag assets according to status. WorkBuddy manages status; UNICUT manages visuals.

I wrote earlier that the most important thing to copy in AI content production is delivery constraints; just making visuals pretty isn't enough. I still hold this judgment. I can tolerate ugly interfaces, but I can't tolerate chaotic versions. What makes me comfortable with WorkBuddy is its ability to turn messy asset lists into filterable, trackable things that others can continue working on. I care less about how fancy its summaries are. Next, I want to try feeding WorkBuddy's exported task cards directly to editing AI to generate rough cuts based on status.