Skip to content

Things to try

You don't need special phrasing. Ask for what you want the way you'd ask an assistant editor, and name projects the way you name them.

Here's what people actually ask for, roughly in the order they discover it.

Where these came from

The reading examples below were run against a real library — seven projects, 250 shots, a mix of YouTube imports and local files. Results are trimmed for length, not invented.

Get your bearings

What's in my StoryFolder library?

You get your projects listed by name, with how long each one is, how many shots it has, and what folder it's in. It's the fastest way to confirm the connection works.

It also knows when a project has lost its footage. If you moved or deleted the original video, the storyboard is still there but the media isn't — and the assistant can see that, so it tells you rather than failing partway through an export later.

Projects in the Trash are left out unless you ask for them.

Have it look at your shots

Show me the first four shots of the makerspace video.

This is the one that changes how the whole thing feels. The assistant doesn't read a description of your shots — it looks at the frames.

It pulls the middle frame of each shot you named and can then tell you what's in them, compare them, spot a continuity problem, or write you a description.

Compare shot 3 and shot 11 — do they cut together?

Which of these shots would work as a thumbnail?

Two things to know:

  • Twelve frames at a time. For a forty-shot video the assistant will work through it in passes.
  • It asks for what you said. You say "shot 3", it fetches shot 3. You never have to think about the numbering.

Tag a batch of shots at once

Go through the hoodie ad and tag each shot with its shot size — wide, medium or closeup.

The assistant looks at the frames, decides, and writes it in. If you don't have a Shot Size field yet, it can create one first.

This is the workflow that pays for itself. Twenty shots of tagging becomes one request, and because it goes in as a single edit, it's a single undo if you don't like the result.

More things in the same shape:

Mark shots 4 through 9 as B-roll.

Favourite every shot with the product in frame.

Add a Location field and fill it in for the shots that are outdoors.

Re-writing a value that's already there does nothing at all — no edit, no history entry, no flicker in the app. So you can ask twice without making a mess.

Find things across everything you've made

Which shots anywhere in my library are marked as favourites?

One search covers every project, not one per project. You get the count, which projects the matches are in, and the shots themselves with their timecodes.

The search reads your notes, not the picture. So it works on whatever you've recorded:

Find every shot tagged Closeup.

Which shots have nothing in the Emotion field?

That last one is the one people end up using most — it's how you find the gaps in a pass you didn't finish. And the answers are usefully different: "nothing matched" and "you haven't filled that field in on anything" don't look the same in the reply.

If you want to search on something you've never recorded, tag it first — see the section above — and it's searchable from then on.

Start an export and carry on working

Export a PDF storyboard of the hoodie ad.

Exports don't hold up the conversation. The assistant starts the job, hands you a reference, and you can keep talking or go back to editing. Ask later:

Did that export finish?

It can produce any of the things StoryFolder exports:

Ask forWhat you get
A storyboardA PDF
ImagesA PNG for every shot
ClipsA separate video file for every shot
A shot listAn XLSX spreadsheet with timecodes
A transcriptSRT and VTT by default; TXT, CSV and Fountain on request
An online storyboardA shareable link

It checks the export actually produced something. A job can finish having written no files at all, and the assistant can tell the difference — so "finished" from it means finished, not just "stopped".

Where exports land

Say where you want them and they go there. Say nothing and they go to:

~/Documents/StoryFolder Exports/<project name>

The assistant can't see your file system, so it doesn't guess. It used to invent plausible-looking folders that didn't exist, and the export would fail halfway through. Now the folder is checked before the job starts, and the assistant tells you the exact path it used.

Import a video

Import ~/Movies/interview-take-3.mov into my Client Work folder.

It checks the file is readable, then starts the import in the background. Ask for progress and it'll check. This is the one case where you do need to give it a real file path — it can't browse your drive.

Keep an eye on what it changed

What have you changed in this project?

Everything the assistant writes goes into StoryFolder's Version history, tagged as made by an assistant rather than by you. So you can ask it to account for itself, and then undo exactly those entries in the app.

Your open StoryFolder window updates live as it works. You don't need to reload anything.

When something's off

StoryFolder isn't open

You get a sentence, not an error:

StoryFolder is not running. These tools work by talking to the StoryFolder desktop app on this computer, so the app has to be open. Open StoryFolder, wait for your library to appear, and try again.

Open the app and ask again. You don't need to restart the assistant.

A note on shot numbers

Every shot has two numbers behind the scenes, and they're off by one: the number you see (shot 1, shot 2…) and the number the tools count from (0, 1…).

You should never have to care. Say "shot 5" and you get shot 5. But if an assistant ever quotes a shot number that's one off from what you're looking at, that's what happened — tell it to use the numbers StoryFolder shows, and it will.

The exact rule

shot_number is 1-based and is what StoryFolder displays and what exported files are named — Images/Shot-1.png. index is 0-based and is what the tools address shots by. Shot 1 is index 0.

The tools accept both, so no conversion is needed. When something does go wrong, the error names both scales, for example:

This project has 15 shot(s), so shot index 99 does not exist. Valid indices are 0-14, which are shot numbers 1-15.

Next step

What you can ask for — the full list of what it's able to do.