Create · moving pictures · updated 22 October 2026

Video

A short clip from words, from a first frame, or between a first and a last frame. The panel shows the exact price for the length, resolution and sound you pick.

See the recipes Models for /video
  • 27video models
  • 1–30sclip length, by model
  • 20credits for the cheapest clip
  • 3style pictures at most

Entry I

Start here

Run /video with a prompt and a model (type to search; the list starts with the recommended ones and shows each one's lowest price). Optional pictures:

  • start_image: the clip starts from this picture.
  • end_image: the clip ends on this one. It needs a start image.
  • ref1 to ref3: a look or character to match, not a frame.

Use frames or style pictures, not both. A few models only animate a picture and say so if you leave out the start image.

The panel then offers the model's lengths, resolutions, aspect ratios and, where the model makes sound, Audio on or off. Credits are taken only when you press Generate. The button shows the price for the settings you picked, and your free 25 credits a day are spent before paid ones. A run that fails is refunded.

Entry II

Writing the prompt

Subject and action

Who or what, and what they do: "an old fisherman hauls a net onto a small boat". Movement is the point of a video; describe it.

Scene and time

Place, time of day, weather. "A grey harbour at dawn, light rain."

Camera

Angle and movement in film words: close-up, low angle, slow dolly in, handheld, aerial pan, static shot.

Look and pace

Lighting, colour, style, and pacing: slow motion, time-lapse, a gentle change over the clip.

Sound in its own sentences

When Audio is on, describe sound effects, background noise and any spoken line in separate sentences.

One shot

A clip is a few seconds. One subject, one action, one camera move. A story of five events comes out rushed.

Entry III

Starting from a picture

When you give a start image, the picture already fixes the scene. Spend the prompt on what changes: what moves, which way, the one thing the subject does, and what the camera does. Restating the picture competes with it. For a first and last frame, keep the framing the same in both so the prompt describes one change.

Some models charge more when they start from a picture; the panel shows it.

Entry IV

Choosing a model

Recommended today, with their lowest price in credits: Alibaba Wan 3.0 (from 100), Seedance 2.5 (from 344), MiniMax H3 Max (from 250), MiniMax H3 (from 650), Seedance 2.0 (from 228), Seedance 2.0 Mini (from 112). The cheapest clip overall is Grok Imagine Video 1.5 Lite from 20 credits; the most expensive, Runway Aleph 2.0, starts at 1400.

Price grows with every second, with resolution, and with sound. Try the idea short and at a low resolution first, then make the final one.

Entry V

Enhance

Enhance has a small chat model (Claude Haiku 5.5, GPT-6 Luna, MiMo v2.6 Flash or DeepSeek V4.1 Flash 1017) rewrite your prompt for the model you picked, for +2 credits. You see both versions and choose Using: Yours or Using: Enhanced, and you can edit the enhanced one before generating.

Entry VI

Every option

What you can type after /video, read from the command itself. You can run it again after 20 seconds.

OptionWhat it doesTakes
prompt needed Video prompt text
model needed Video model (type to search) texttype to search
start_image Picture the video starts from (optional) file
end_image Picture the video ends on (needs start_image) file
ref1 … ref3 Style or character look (not a start/end frame) file
hidden Only you can see this (default: false — post in channel) true / false

Entry VII

Recipes

Requests you can copy into Discord and change to fit.

A calm establishing shot

Copy this

/video prompt: A lighthouse on a rocky coast at dusk, waves breaking below, the beam starting to turn. Slow aerial pan from left to right. Soft orange and blue light, cinematic. model: (a recommended model)

Why it works. One place, one movement, one camera move. Pick 16:9 in the panel.

Bring a photo to life

Copy this

/video prompt: Her hair and scarf move in the wind, she turns her head to the camera and smiles. The camera holds still. start_image: (your photo)

Why it works. With a start image, describe only what changes.

With sound

Copy this

/video prompt: A barista pours steamed milk into a cup, making a leaf pattern. Close-up, shallow focus. Sound: the hiss of the steam wand, cups clinking, quiet café chatter.

Why it works. Sound is described apart from the picture. Turn Audio on in the panel on a model that has it.

Entry VIII

Tips

  1. I

    Make a short, low-resolution test first. Length and resolution decide most of the price.

  2. II

    Say the camera move in film words; models know "dolly in" and "pan left".

  3. III

    Describe the things you want instead of listing what you do not want.

  4. IV

    Vertical for phones: pick 9:16 in the panel when the model offers it.

Entry IX

Limits

  • One video at a time per person, and 20 seconds between /video commands.

  • A clip that is not ready within 15 minutes is stopped and refunded.

  • Frames and style pictures cannot be mixed in one run, and an end frame needs a start frame.

  • Lengths, resolutions and aspect ratios are only those the chosen model supports.

Entry X

Questions

Why does the same prompt cost different amounts?

Video is priced per second, and resolution, sound and sometimes a start picture change the rate. The Generate button shows the total.

The aspect menu disappeared

Some models take the shape from your start picture, so the panel hides the aspect choice.

Can I make it longer later?

No. Each run is one clip. Pick the length in the panel before you generate.

Can I upscale a clip?

Yes, with /clip upscale.

/clipClip /imageImage /wikiWiki

Where the advice comes from

Prompting advice is from these free public guides, matched to what ALMA lets you set. Prices and limits are read from the bot itself, as of 22 October 2026.

Try it

Type /video in any server with ALMA, or in a direct message to the bot.

All guides

Something here wrong, or missing? Tell us.