Is AIR Live actually streaming generated video?

Yes in Live Director mode. fal H3 Max Director sends continuous video and audio over WebRTC and accepts new prompts during the session. Preview mode creates finite five-second clips and is labelled separately.

How does voice direction work?

Browser speech recognition transcribes your words, and Evolink turns completed sentences into scene directions. A configured Gemini Live connection can handle speech instead. The generated video is saved to your account; AIR Live does not store raw microphone audio.

Why does one Live session cost 72 credits?

fal bills Director sessions for at least 60 seconds. The launch credit price uses the announced post-promotion rate, plus speech, storage, payment fees, and operating margin.

Is this an official fal or MiniMax product?

No. AIR Live is an independent product that currently uses fal-hosted MiniMax H3 Max Director as its video engine.

Real-Time AI Video Generator & Voice Director

Your first real-time AI video generator session

What can a real-time AI video generator make?

Start with a red tram crossing a wet street, a window opening onto the sea, or a person entering an empty station. A real-time AI video generator needs a recognizable opening. The real-time AI video generator can then develop what happens next.

AIR Live is a real-time AI video generator for personal creation. Your session belongs to your account; strangers do not vote on your next shot. Keep one subject and one setting at first, so the real-time AI video generator output is easy to judge.

How does a real-time AI video generator use speech?

The real-time AI video generator receives a scene direction after speech recognition and text cleanup. Saying “keep the tram, move closer” gives the real-time AI video generator a smaller, clearer change than repeating every detail of the entire street.

Choose the recognition language before speaking. A real-time AI video generator cannot correct words it never heard, so check the real-time AI video generator transcript when a name sounds wrong. English pages default to English; Chinese pages default to Chinese. Both remain manually selectable.

Does a real-time AI video generator start instantly?

No. A real-time AI video generator still needs an initial connection and a first generated frame. In one controlled test, our real-time AI video generator received its first usable picture after roughly eleven seconds. That observation is not a speed guarantee.

During playback, a real-time AI video generator may already have prepared footage ahead of the visible frame. A new direction changes upcoming content, not the picture you have already watched. Give the scene time to respond before reversing the instruction.

Keep a real-time AI video generator scene consistent

Tell the real-time AI video generator what must stay: the same person, the red coat, the station platform. Then name the change. The real-time AI video generator has more useful guidance from “move behind her” than from “make this more cinematic.”

A real-time AI video generator can also make a deliberate transition. Ask for a cut, a gradual change of light, or a reveal beyond a doorway. Those are different choices. Precise scene language matters more than a long list of camera brands.

Choose a real-time AI video generator mode

Use the real-time AI video generator in Live mode when your next decision depends on the current picture. The real-time AI video generator also offers a separate short-video mode. That produces a five-second file, not a continuous directing session.

With this real-time AI video generator, speech in short-video mode fills the editable scene description. The Generate action submits the clip. An unavailable engine is marked unavailable; clicking does not manufacture a new result from the example video.

Where does the real-time AI video generator save results?

A real-time AI video generator should leave you a usable recording. AIR Live records generated picture and sound, uploads the file to cloud storage, and associates it with your account. The real-time AI video generator keeps that output separate from your microphone input.

Wait for the real-time AI video generator to finish saving before closing the tab. If an upload fails, the available footage has a retry path. A stopped stream and a saved file are different states; check the recording before starting another take.

What does the real-time AI video generator cost?

The real-time AI video generator uses 72 credits for a sixty-second Live session and 4 credits for a generated five-second clip. Check the selected mode first. The real-time AI video generator shows the cost before the action that starts generation.

Stopping a real-time AI video generator session early does not remove the supplier’s one-minute minimum. A free trial would therefore need a site-funded allowance. Viewing an example, reading a transcript, or replaying an existing video is not another paid generation.

Review a real-time AI video generator take

Give the real-time AI video generator a wide opening shot. Once the subject is recognizable, request one camera move and then one lighting change. This lets you assess the real-time AI video generator without mixing five unrelated ideas into a single test.

Review the real-time AI video generator recording after saving it. Look for the moment the movement changes, whether the subject stays recognizable, and where you would cut the sequence. Keep those observations for the next take instead of relying on a general impression.

Real-Time AI Video Generator

Speak a scene. Then change what happens next.

Speak

Direct

Stay live

Preview cheaply. Go live when the idea is worth a minute.

Spark

Studio

Broadcast

What “real-time video” means here.