Director Takes Calls: Start a Project by Talking to It
Good ideas rarely wait until you are sitting down. Now you can hand one over before it goes.
An idea turns up while you are walking, or on the sofa, or in the ten minutes between two meetings. By the time you are back at a desk it has gone soft, and what you write down is a tidier, smaller version of the thing you actually had.
Director takes calls now. The wave icon beside the send arrow starts one, it runs in a phone browser, and it means the idea can be handed over while you still have all of it.
One unfinished sentence
The film above is a real session. It opens with somebody saying this, out loud, to an empty canvas:
I’m thinking of building out an idea for an explainer video with 3D cutouts, cardboard, and construction paper. I want to understand the history behind cinema, and maybe make it a 30-second short film.
That is not a brief. It is thinking out loud, with a “maybe” in it. Written down it would have been pruned into something safer on the way to the keyboard.
Director took it and started working.
What happens on the other end
The call is not dictation. Director is not waiting for you to finish so it can go away and act.
Inside that one conversation it wrote a plan, generated six reference images for the milestones it picked out of cinema history, edited one of them when asked, tidied the canvas into groups, built the keyframes, cut a 30-second film and generated a voiceover for it.
The project it handed back is called Cinema’s Daring Gambles. Lumière in 1895, Méliès and the Moon in 1902, pictures learning to talk in 1927, colour in 1939, Star Wars building a studio from scratch in 1977, Toy Story going fully computer-animated in 1995. All of it in paper and cardboard, because that was in the sentence.
None of those six milestones were specified. They were the answer to “some interesting movies that made a bit of cinema history”.
The bits of the call that matter
It shows you what it heard. Every spoken request lands in the thread as text, tagged Via voice. You can see the sentence it actually parsed, not just the result.
It is a bar, not a mode. While the call runs you get a small bar showing Listening or Speaking with a timer, a mute button, and a red button to end it. The rest of the app carries on underneath.
You can still type. The text box stays live below the call bar. Anything awkward to say out loud, a file name, a pasted reference, goes in by hand without hanging up.
The call is kept. When it ends it collapses into a single line in the thread, Voice call 1:14 · 6 turns. Nothing said on a call is lost.
It pushes back
This is the part that makes it usable rather than a party trick.
Asked for more keyframes, Director did not make them. It said the keyframes would be built from references that had not been approved yet, so it was holding them until they had been looked at, and asked whether any prop needed changing first.
When it was asked to group the canvas it did it, then explained what it could not do and why, because an item cannot belong to two groups at once.
And when the film was finished it said this:
Your 30-second video is ready. I haven’t watched it yet, so I can’t say how the cuts, card text or sound turned out.
An assistant that tells you the limits of its own confidence is worth more than one that tells you everything went well.
Where typing still wins
Voice is for the shapeless end of a job. Describing a look, changing your mind, saying “no, warmer” without composing a sentence first.
Typing is better the moment precision beats speed. Exact names, pasted links, and the prompt craft the manual panel rewards are all easier written than spoken.
The two are not a choice. The call bar and the text box are on screen at the same time, and most real sessions use both.
The short version. The hard part of a project is usually the first ten minutes, and they tend to happen somewhere you cannot type. Open the project, press the wave icon, and say the half-formed version. Director starts building while you are still talking.
Calls are available on paid plans. The rest of our tutorials are inside MITO too, under Tutorials in the sidebar.
Frequently asked
- How do I talk to Director?
The wave icon sits beside the send arrow in the Director panel. Tap it and the call starts. A bar appears showing whether it is listening or speaking, with a timer, a mute button and a red button to hang up. Calls are available on paid plans.
- Can I use Director voice on my phone?
Yes. It runs in a phone browser, so you can open a project and start talking without being at a desk. That is the point of it: the idea gets handed over while you still have it.
- Does Director keep working while I am talking?
Yes. It is not taking dictation and acting later. During the call in the film it wrote a plan, generated six reference images, edited one of them, grouped the canvas, built keyframes, assembled a 30-second cut and generated a voiceover.
- Can I still type during a call?
Yes. The text box stays live underneath the call bar, so you can paste a link or type a name that would be awkward to say out loud, without hanging up.
- Is there a record of what I said?
Every spoken request lands in the thread as text, tagged Via voice, and the call itself is saved as a collapsible entry showing its length and how many turns it took. Nothing said on a call disappears.
- Does it just do whatever I say?
No, and that is the useful part. In the film it holds the keyframes back because they would be built from references you have not approved yet, and it says so. It also tells you plainly when it cannot judge its own output.
- When is typing still better?
When precision matters more than speed. Exact names, pasted references, and the prompt craft the manual panel rewards are all easier typed. Voice is for the part where you are still working out what you want.
- What can I actually ask for?
Anything you would ask a person. A look, a format, a length, a fix to one specific thing. The session in the film starts from one unfinished sentence about cardboard and cinema history and ends with a finished 30-second film.
The AI tool for cinematic video.
Generate, direct, and publish professional videos — powered by the best AI models.
More posts
