My Business Growth Kit

Article

How to Make a Short Film With AI (2026): A Complete 60-Second Project

November 4, 2026 · 14 min read

How to Make a Short Film With AI (2026): A Complete 60-Second Project

Knowing how to make a short film with AI is mostly about planning, not about finding a magic tool. Most beginners generate a few impressive clips, then discover that the hero has a different face in every shot and the harbour changes colour between cuts. This guide fixes that by walking you through one complete 60-second project from idea to export. You will get a ready-to-use script, a ten-shot list, copy-and-adapt prompts, voiceover and sound advice, editing steps and a troubleshooting table for the consistency problems that trip people up. It works with any current image or video generator, because the method matters more than the brand.

Key takeaways

  • Treat 60 seconds as ten shots of about six seconds each, with one character, one location and one visible change.
  • Consistency is created before you generate: write a character sheet, a style line and a location description, then paste them into every prompt.
  • An image-first workflow (make a still, then animate it) usually gives beginners more control than pure text-to-video.
  • Expect to generate several takes per shot, and plan to cut around problems you cannot fix.
  • Check each tool's licence, voice-consent rules and any AI-disclosure requirements in your country or platform before you publish.

Plan a 60-second AI short film before you open any tool

A short film made with AI fails or succeeds at the planning stage. Generators are good at single beautiful moments and weak at remembering what they did last time, so your plan has to carry the continuity.

Choose one idea you can tell in ten shots

Pick a story with one main character, one location and one change. "A lighthouse keeper lights a lantern and a child notices it" works. "A detective chases a thief across five cities" does not, because every new location is a new consistency problem. Write the idea as a single sentence: who, where, what changes.

Set a timing and generation budget

Many generators produce clips of roughly 4 to 10 seconds, so ten shots of about 6 seconds fills a minute. Plan on three to five attempts per shot, which means 30 to 50 generations. Check how your chosen tool counts credits or usage before you start so you are not surprised halfway through. Our guide on what a short film made with AI video actually costs breaks down where the time and spending usually go.

Pick image-first or text-to-video

In an image-first workflow you generate a still frame for each shot, approve it, then animate it. In text-to-video you describe the shot and receive a moving clip directly. Image-first costs a little more effort but lets you reject a wrong face or costume before spending on motion. If you are still choosing software, see the comparison of AI video generators for filmmakers by control, consistency, length and licensing.

The example project: script and shot list for a 60-second film

Our example is called "The Lantern Keeper". It is deliberately simple: one older woman, one small harbour, one lantern, and a child at the end. Swap in your own idea but keep the structure.

Logline and voiceover script

Logline: At dusk, a harbour keeper lights a paper lantern and sets it adrift, and a child on a distant jetty sees it.

Voiceover (about 35 words, spoken slowly with pauses): "Every evening, Mara lights one lantern. Not for the ships. For whoever needs finding. The harbour has forgotten many things. It has never forgotten the light."

Keeping narration short leaves room for ambient sound and lets the pictures do the work, which also hides small visual flaws.

The ten-shot list

ShotTimeFramingActionSound
10-6sWideHarbour at blue hour, boats at restGulls, water
26-12sMediumMara on the pier holds an unlit paper lanternVoiceover line 1
312-18sClose-upHands strike a match, lantern glowsMatch scrape
418-24sClose-upWarm light on her faceVoiceover line 2
524-30sWideShe lowers the lantern onto a small raftWater lapping
630-36sLow angleRaft drifts, light reflects on waterVoiceover line 3
736-42sWideRaft passes moored boatsSoft music rises
842-48sMediumChild on a distant jetty notices the glowVoiceover line 4
948-54sClose-upChild's face lit, reaching outMusic
1054-60sWideMara in silhouette, tiny light far awayFade out

If two characters feel risky, show the child only as a small silhouette in shots 8 and 9. That removes the hardest consistency task.

Build a consistency kit so your characters keep their faces

The most searched problem in this field is how to keep characters consistent in AI video. The honest answer is that no prompt guarantees it, but a written kit improves your odds sharply.

Write a character sheet

Describe your hero in fixed wording and reuse it word for word. Example: "Mara, a woman in her 60s, weathered tan skin, silver hair in a long braid over her left shoulder, round tortoiseshell glasses, mustard-yellow waterproof raincoat with the hood down, navy wool scarf." Include three or four distinctive, easy-to-see features. Vague words like "kind old lady" invite the model to improvise.

Lock the style and the location

Create a style line and paste it into every prompt: "cinematic 35mm film look, soft blue-hour light, muted teal and amber palette, light film grain, shallow depth of field." Do the same for the setting: "small stone harbour, wooden pier, a few painted fishing boats, calm dark water." If your tool supports reference images, character references or seed values, use them, but features and names vary by product, so check its documentation.

Use a reusable prompt template

  1. Shot type and camera angle
  2. Character block (copied exactly)
  3. Action, in one simple verb phrase
  4. Location block (copied exactly)
  5. Lighting and time of day
  6. Style line (copied exactly)
  7. Exclusions, such as "no text, no logos, no extra people"

Save these blocks in a document so you paste rather than retype. Small wording changes are a leading cause of drift.

Prompts for each shot: image prompts and motion prompts

Write the still image prompt first, then a separate short motion prompt. Keeping the two apart makes it easier to see which one caused a problem.

Example image prompts

  • Shot 2 (medium): "Medium shot, eye level. Mara, a woman in her 60s, silver braid over her left shoulder, round tortoiseshell glasses, mustard-yellow raincoat, navy scarf, holds an unlit paper lantern. Wooden pier, small stone harbour, painted fishing boats, calm dark water. Blue hour. Cinematic 35mm film look, muted teal and amber, light grain, shallow depth of field. No text, no extra people."
  • Shot 3 (close-up): "Close-up of weathered hands in navy wool sleeves striking a match beside a paper lantern, warm glow beginning inside the paper. Blue-hour harbour blurred behind. Same style line."
  • Shot 6 (low angle): "Low angle at water level, a small wooden raft carrying a glowing paper lantern drifting on calm dark water, warm reflection, moored painted boats softly out of focus. Same style line."
  • Shot 8 (medium): "Medium shot of a child in a red knitted hat on a distant wooden jetty, looking toward a faint warm light on the water. Same harbour, same blue hour, same style line."

Motion prompts that avoid weird movement

Ask for one movement at a time and keep it gentle: "slow push-in, the lantern flame flickers slightly, water ripples softly, character stays still apart from a slow blink." Fast action, crowds, and complex hand movement are where models break most. Use "static camera" when you want to reduce wobble. For a wider view of the whole workflow, our AI filmmaking for beginners pipeline shows how these steps connect from script to export.

Troubleshooting AI video consistency problems

Expect problems. Diagnosing them quickly is the real skill. Work through the table from the cheapest fix to the most expensive.

Problem, cause and fix table

ProblemLikely causeFix
Face changes between shotsCharacter described loosely or differently each timePaste the exact character block; use a reference image if your tool allows; keep faces smaller in wide shots
Coat colour or props driftToo many details, or a lighting changeReduce to three key items; name colours explicitly; repeat them in every prompt
Harbour looks different each shotLocation block rewordedReuse the exact location text; generate wide shot first and use it as a reference
Style flickers (grainy, then clean)Style line missing or alteredPaste the same style line; add grain and colour in editing for uniformity
Hands or lantern warpComplex motion in the promptSimplify the action; use a close-up still with a tiny movement; cut away early
Slow-motion or floaty lookOver-long clip or vague motion wordsDescribe specific, small movement; trim the clip; speed up 5 to 10 percent in editing
Unwanted text or extra peopleModel filling empty spaceAdd exclusions; crop; regenerate
Two characters blendBoth described in one promptGenerate them separately; composite or cut between them

When to cut around a problem instead of fixing it

Professionals do this constantly. If a take is almost right, use only the clean three seconds. Trim the first and last half-second of every clip, because artefacts often appear at the edges. Replace a failing close-up with a silhouette, a hand, or a reflection. Stay alert for the diminishing returns trap: if five attempts fail, rewrite the shot rather than rolling again.

AI voiceover and sound design for your film

Sound hides visual imperfections and makes a film feel finished. Many beginners spend nearly all their time on pictures and then add silence, which is why their clips feel like demos.

Write narration for the ear

Read every line aloud. Use short sentences, and mark pauses with full stops. At a slow documentary pace, 60 seconds holds only around 80 to 100 words at most, and fewer if you want breathing room. Our example uses about 35.

Generate and check the voice

When using AI voiceover for film, generate each line as a separate file so you can redo one without redoing everything. Listen for mispronounced names and odd stress, and adjust spelling phonetically if needed ("Mah-rah"). Only use voices you have the right to use. Cloning a real person's voice without their permission is unsafe legally and ethically, and the rules vary by country and by tool.

Layer ambience, effects and music

  1. Ambience: a continuous harbour bed under the entire film.
  2. Effects: match scrape, water lapping, footsteps, timed to the picture.
  3. Music: a quiet bed that rises in shots 7 to 9. Use only music you are licensed to use.
  4. Levels: keep narration clearly above music. Many platforms adjust loudness automatically, so aim for steady levels rather than extremes.

AI film editing, colour matching and export

Editing is where several mismatched clips become one film. Any ordinary video editor works, and some include AI features such as auto-captions or noise reduction, though the basics are the same.

Assemble the cut

Place the voiceover first, then lay clips against it. Cut on movement where possible, and hold most shots for four to six seconds. Vary your framing: wide, medium, close-up, so the eye does not notice small face differences. Add a one-second fade at the beginning and end.

Match colour and grain

Apply the same colour adjustment layer across all clips. Nudge shadows toward teal and highlights toward amber to unify the palette, then add one consistent layer of film grain over the whole timeline. This single step hides much of the style flicker that generators produce.

Export settings

For general sharing, export 1920 x 1080 pixels (1080p) at 24, 25 or 30 frames per second, matching your source clips, as an H.264 MP4. For vertical platforms, use 1080 x 1920. Watch the finished film on a phone and on a laptop with headphones, then fix any distracting moments before publishing. Add captions, which help viewers in noisy places and support accessibility.

Rights, disclosure and realistic costs

Making an AI generated short film raises practical questions beyond the creative ones. Answer them before you publish, not after.

Check licences and platform rules

Each tool sets its own terms about commercial use, ownership of outputs and use of likenesses, and these terms can change. Read them for the plan you use and keep a record. Rules on labelling AI-generated content differ between countries and platforms, and some are still developing, so check current requirements where you publish and where your audience lives. Avoid prompts that imitate living artists or real people. If you use AI to draft scripts or research, verify facts, as covered in how to check AI output when you are not the expert.

Plan time and money honestly

Expect a first 60-second film to take several evenings, not one. Most of that time goes on repeat generations, audio and cleanup. Costs depend on the tools you choose and your local currency, so compare current plans on each provider's own site.

If you want a structured route, the AI Filmmaker Blueprint from My Business Growth Kit combines ebooks, audiobooks, AI tutor prompts, workflows and exercises for beginners. It is a digital download you read or listen to in your online library. Details are on the product page, and no outcome is promised: your results depend on practice, tools and effort.

Conclusion

A 60-second film is manageable when you keep it to one character, one location and ten shots. Write your character, style and location blocks once and reuse them exactly. Generate stills first, animate gently, and treat every failed take as information. Use the troubleshooting table to diagnose problems, cut around what you cannot fix, and unify the film with consistent colour, grain and sound. Finally, check licences and disclosure rules before publishing, then start your own first minute.

Frequently asked questions

How long does it take to make a short film with AI?

A first 60-second film usually takes several sessions, often a few evenings. Most of the time goes on regenerating shots, fixing audio and editing rather than writing prompts. Your pace will improve once you have reusable character and style blocks.

Do I need to know how to edit video to make an AI short film?

You need basic editing skills: placing clips on a timeline, trimming, adding audio and exporting. Any entry-level video editor can do this, and you can learn the essentials in a few hours. AI features in some editors can help with captions or noise reduction.

How do I keep characters consistent in AI video?

Write a fixed character description and paste it unchanged into every prompt, with three or four distinctive features. Use reference images or character features if your tool offers them, and generate stills before animating. Perfect consistency is not guaranteed, so use varied framing and cut around flawed takes.

Should I use image-to-video or text-to-video?

Beginners often get better control with image-to-video, because you approve the still frame first and then animate it. Text-to-video is faster for quick experiments but gives you less say over the face, costume and setting. Many filmmakers use a mix.

Can I use AI voiceover in a film I publish?

Often yes, but check the voice tool's licence for commercial use and its rules on consent. Never clone a real person's voice without their permission. Regulations and platform policies differ between countries, so check where you publish.

How many shots does a 60-second film need?

Around ten shots of roughly six seconds each is a practical target for beginners. It gives enough variety without creating too many continuity problems. Shorter, simpler shots are easier to keep consistent.

Do I have to tell viewers my film was made with AI?

Requirements vary by country, platform and use, and some rules are still evolving. Many platforms ask creators to label realistic AI-generated content. Check the current policy on each platform where you publish, and consider being open about your process anyway.

What export settings should I use for an AI short film?

For most sharing, use 1920 x 1080 (1080p) at 24, 25 or 30 frames per second as an H.264 MP4, matching your source clips. For vertical platforms use 1080 x 1920. Always test playback on a phone before publishing.

All articles