Browse Library
Free Prompt Library

30 Extreme Close-Up Prompts

ECU and macro prompts crafted by a VFX artist — eyes, micro-expressions, hands and texture detail. Paste into Kling AI, Runway or Sora.

30
Prompts
5
Categories
100%
Free
Sample AI generated image for prompt library

About These Extreme Close-Up Prompts

The extreme close-up is the most intimate shot in cinema — a single eye, a hand tightening, a key entering a lock. Directors cut to an ECU when a detail carries more story than a whole room. It is also the shot type where AI video performs at its best: with all of the model’s resolution spent on one small region, a trembling lip or a droplet of condensation renders with a realism that wide shots can never reach.

But ECUs fail for a specific reason when prompted casually: AI models over-smooth. Ask for "a close-up of a face" and you get plastic skin and dead eyes. Every prompt in this collection fights that with explicit texture language — visible pores, fine vellus hair, thread-by-thread detail — plus a macro lens specification (85–100mm) and exactly one micro-action per shot, because a single clean movement is what keeps extreme close-ups from warping.

These 30 prompts cover the five jobs an ECU does in a real edit: eyes, micro-expressions, hands, tension-building object inserts, and pure texture. Use them as cutaways inside a larger scene, or as standalone shots — every one specifies lens, movement, lighting direction and grade.

How to Use These Prompts

1
Copy any prompt
Browse the collection and find a prompt that matches the type of shot you need. Click the Copy button to copy the full prompt to your clipboard.
2
Open your AI video tool
Open Kling AI, Runway, Sora, or any other AI video generator you use. Navigate to the text prompt input field.
3
Paste and adjust
Paste the prompt into the input field. You can use it exactly as written, or replace the generic scene description with your specific subject — for example, replace "a lone warrior" with "a detective in a rainy alley".
4
Generate and review
Generate the video clip. If the result does not match what you expected, try adjusting the lighting description or camera movement in the prompt and generate again.

Tips for Best Results

One micro-action per shot — "a single slow blink" holds together; "she blinks, smiles and looks away" will warp. Save the second action for the next generation.
Fight the smoothing: add "visible skin texture, fine pores, no digital smoothing" to any face ECU. Without it, AI models default to plastic beauty-filter skin.
Use 85–100mm macro lens language and name the shallow focus — "brow and cheek dissolving into soft blur" tells the model exactly where detail should live and where it should melt.
Describe emotions as muscle movements, not labels — see the expression table below. "Jaw sets, masseter flexes" performs; "angry face" doesn’t.

Writing Expressions AI Can Actually Perform

Emotion labels like "sad face" or "angry expression" produce generic, often uncanny results — the model has no idea which muscles you mean. The fix is the same direction real actors are given: describe the physical behaviour, not the feeling. In an extreme close-up the model spends all of its resolution on one region of the face, so a trembling lip or a two-millimetre brow movement renders better here than in any other shot type.

Emotion✗ Don't write✓ Write instead
Griefa sad expressioneyes glassy with unshed tears, lower lip pressed tight, one slow blink
Feara scared facea hard swallow, larynx rising and falling, chin trembling almost invisibly
Suppressed angerlooking angryjaw setting, masseter flexing once, lips pressing into a thin line
Reliefa relieved smilethe brow unknotting, eyes closing for one long second, a small exhale
Suspiciona suspicious lookeyes narrowing a few millimetres, brow lowering slowly, gaze holding still

Frequently Asked Questions

What is an extreme close-up (ECU) shot?
An extreme close-up frames a single small detail — an eye, a mouth, a hand, an object — so it fills the entire frame. In film grammar it is used for maximum emotional intensity or to make the audience notice a story-critical detail. In prompts, it is signalled with "extreme close-up" plus a long macro lens such as 100mm.
Why do AI extreme close-ups often look plastic or fake?
Because most AI video models are trained toward flattering, smoothed results. At ECU distance that smoothing becomes obvious — skin turns to wax. The fix is explicit anti-smoothing language in the prompt: visible pores, fine vellus hair, natural skin texture, film grain. Every face prompt in this collection includes it.
Can these ECU prompts show emotions and expressions?
Yes — but only if the emotion is written as physical behaviour rather than a label. "Sad expression" produces generic results; "eyes glassy with unshed tears, lower lip pressed tight" gives the model concrete muscle geometry to render. The Micro-Expressions category and the table on this page are built entirely around that principle.
How do I use these inside a full scene with other shot types?
ECUs work best as inserts inside a wider sequence — a scene plays in wide and medium shots, then cuts to an extreme close-up at the emotional peak. Our free AI Scene Builder includes an Insert Shot type and keeps lighting and characters consistent between your wide shots and these close-ups.

Build ECUs into a full consistent scene

Scene Builder generates matched wide shots, close-ups and inserts with locked characters and lighting.

🎥 Open AI Scene Builder