ControlNet Poses Reference
Curated pose descriptions and keywords for AI image generation with ControlNet
Tip: These pose keywords work best when combined with ControlNet OpenPose or used as descriptors in any AI image generator. For best results with ControlNet, use a reference image with the pose preprocessor.
Standing Neutral
Basic standing position with relaxed arms
standing, neutral pose, arms at sides, facing forward
Standing Confident
Assertive pose with hands on hips
standing, confident pose, hands on hips, power stance
Standing Casual
Relaxed standing with hand in pocket
standing, casual pose, one hand in pocket, relaxed
Contrapposto
Classical weight shift pose
contrapposto, classical pose, weight on one leg, S-curve
Arms Crossed
Standing with folded arms
standing, arms crossed, defensive pose, closed body language
Looking Back
Standing facing away, looking back
standing, looking over shoulder, turned head, profile view
Sitting Upright
Formal seated position
sitting, chair, upright posture, hands on lap
Sitting Casual
Relaxed seated position
sitting, relaxed, leaning back, casual pose
Sitting Cross-legged
Cross-legged on floor
sitting, cross-legged, floor, meditation pose
Sitting Side
Elegant side-seated pose
sitting, side view, legs together, elegant pose
Sitting Forward
Leaning forward engaged pose
sitting, leaning forward, elbows on knees, engaged
Sitting Legs Crossed
Professional crossed legs seated
sitting, legs crossed, elegant, professional
Walking
Mid-stride walking pose
walking, mid-stride, one foot forward, dynamic
Running
Athletic running pose
running, sprinting, dynamic motion, athletic
Jumping
Mid-air jumping pose
jumping, mid-air, legs bent, dynamic leap
Dancing
Dynamic dancing pose
dancing, graceful movement, arms extended, flowing
Reaching Up
Reaching toward sky
reaching up, arms raised, stretching, looking up
Throwing
Throwing motion pose
throwing, arm back, dynamic motion, athletic pose
Head Tilt
Slight head tilt for portraits
portrait, head tilt, soft expression, three-quarter view
Chin Down
Chin lowered, eyes up
portrait, chin down, looking up, mysterious
Profile
Pure side profile view
profile view, side view, silhouette, one eye visible
Three-Quarter
Classic portrait angle
three-quarter view, angled face, classic portrait
Looking Away
Gaze directed away from camera
looking away, candid, thoughtful, averted gaze
Direct Gaze
Direct eye contact with viewer
direct eye contact, facing camera, intense gaze, frontal
Hands Together
Clasped hands gesture
hands together, clasped hands, praying position, fingers interlaced
Hand on Face
Hand resting on chin/cheek
hand on face, chin rest, thoughtful, contemplative
Pointing
Pointing gesture
pointing, index finger extended, gesture, indicating
Open Palms
Open palm gesture
open palms, welcoming gesture, hands up, peaceful
Fist
Closed fist pose
clenched fist, power gesture, strength, determination
Reaching Out
Hand reaching toward viewer
hand reaching, extended arm, offering hand, connection
Lying Down
Lying flat on back
lying down, supine, on back, relaxed
On Side
Reclining on side
lying on side, reclining, curved pose, relaxed
Propped Up
Propped on elbows
lying, propped up on elbows, casual, relaxed
Curled Up
Curled comfortable position
curled up, fetal position, cozy, resting
Showing 34 of 34 poses
Related Tools
Prompt Enhancer
Expand simple ideas into detailed, descriptive prompts for better image generation
Prompt Weight Calculator
Calculate and format attention weights for Stable Diffusion (()) and Midjourney (::)
AI Resolution Guide
Reference guide for optimal resolutions and upscaling factors
Stable Diffusion Prompt Builder
Build prompts with support for LoRAs, embeddings, and common checkpoint styles
AI Art Style Reference
Visual library of artist styles, mediums, and lighting modifiers
Aspect Ratio Calculator
Calculate dimensions and verify aspect ratios for Midjourney, DALL-E, and Stable Diffusion
What is ControlNet?
ControlNet is a neural network extension for Stable Diffusion that adds precise control over image generation. It uses reference images to guide the AI, allowing you to maintain specific poses, compositions, or edges while generating new content.
The OpenPose preprocessor extracts skeleton/pose information from reference images, letting you transfer poses to AI-generated characters.
How to Use These Poses
With Text Prompts
Copy the keywords and add them to your prompt. Works with any AI image generator — Midjourney, DALL-E, or SD.
With ControlNet
Find a reference photo matching the pose, run it through the OpenPose preprocessor, and use the resulting skeleton as your ControlNet input.
Combine Approaches
Use both pose keywords in your prompt AND a ControlNet reference for maximum control over the final result.
ControlNet Preprocessors
OpenPose
Extracts body skeleton and keypoints. Best for full-body poses and character positioning.
Depth
Creates depth maps for 3D positioning. Good for maintaining spatial relationships.
Canny
Edge detection for precise outlines. Great for maintaining exact shapes and compositions.
MediaPipe Face
Facial landmark detection. Perfect for portrait expression and face angle control.
Pro Tip: Reference Images
Stock photo sites like Unsplash and Pexels are great sources for pose reference images. Search for the pose name to find photos you can run through ControlNet preprocessors.
Frequently Asked Questions
Do I need ControlNet to use these poses?
No! The pose keywords work with any AI image generator. ControlNet provides more precise control, but text prompts alone can produce good results for basic poses.
How accurate are text-only pose prompts?
Text prompts work well for common poses but may need multiple generations for complex poses. ControlNet with a reference image gives much more reliable results for specific poses.
Can I use multiple poses together?
For multiple characters, describe each pose separately in your prompt. With ControlNet, you can include multiple pose skeletons in your reference image for group compositions.
Related Tools
Stable Diffusion Builder
Build complete SD prompts with pose keywords.
Style Reference
Combine pose keywords with art styles.
Prompt Weights
Adjust emphasis on pose elements.
Negative Prompts
Avoid bad anatomy in pose generation.