The
in Melies turns your images and text prompts into video clips. AI image to video is the fastest way to bring storyboard frames, photos, and AI-generated images to life. With 8 video models from Google, Kuaishou, ByteDance, MiniMax, Alibaba, and Lightricks, you can create everything from quick test clips to high-quality cinematic footage.This guide walks you through the entire process, from choosing a model to extending and editing your videos.
Quick answer: To turn an AI image to video with Melies, choose from 8 video models (LTX 2 Pro at 50 credits to Veo 3.1 at 400 credits), select image-to-video or text-to-video mode, write a motion prompt describing how the scene should move, and click Generate. For most filmmaking work, use image-to-video with Kling v3 Pro (100 credits, HD, up to 15 seconds) for the best balance of quality and cost.
Text-to-Video vs. Image-to-Video
Melies supports two generation modes:
Text-to-Video
Write a prompt describing a scene, and the AI generates a video from scratch. The model handles everything - composition, characters, motion, and camera movement.
Best for: Quick concepts, abstract scenes, situations where you do not have a specific starting image in mind.
Image-to-Video
Start with an AI-generated image (or upload your own) and the model animates it. You write a motion prompt describing how the scene should move.
Best for: Most filmmaking work. You already created a storyboard frame with the
, and now you want to bring it to life. The result is more predictable because the model starts from your specific visual.Recommendation: Image-to-video is the preferred workflow for filmmakers. Generate your perfect frame first using the image tools (with



Step 1: Choose Your Video Model
Melies offers 8 video models. Here is how to pick the right one:
Premium Quality
Veo 3.1 (400 credits) - Google
- Resolution: 4K
- Duration: 4-8 seconds
- Special: Native audio generation (the video comes with sound)
- Best for: Final production clips where quality is everything
Kling v3 Pro (100 credits) - Kuaishou
- Resolution: HD
- Duration: 3-15 seconds
- Special: Multi-shot support, voice ID
- Best for: The best balance of quality, duration, and features
Mid-Range
Kling O3 Standard (80 credits) - Kuaishou
- Resolution: HD
- Duration: 3-15 seconds
- Special: Character elements
- Best for: Good quality with character consistency features
Seedance v1 Pro (80 credits) - ByteDance
- Resolution: 1080p
- Duration: 2-12 seconds
- Best for: Reliable quality from ByteDance at a mid-range price
Hailuo 02 (70 credits) - MiniMax
- Resolution: 512P or 768P
- Duration: 6 or 10 seconds
- Best for: Good quality for social media and web content
Budget-Friendly
Kling v3 Standard (60 credits) - Kuaishou
- Resolution: HD
- Duration: 3-15 seconds
- Special: Character elements
- Best for: Longer clips at a lower cost, character work
WAN v2.2 (60 credits) - Alibaba
- Resolution: 480-720p
- Duration: ~5 seconds
- Best for: Budget generation, quick concepts
LTX 2 Pro (50 credits) - Lightricks
- Resolution: HD
- Duration: ~4.8 seconds
- Special: Camera LoRA for precise camera movement control
- Best for: Fast generation with specific camera movements
Step 2: Write Your Video Prompt
Video prompts are different from image prompts. You need to describe motion, not just a static scene.
Prompt rewrite assistant: The prompt field has a magic wand icon in the top right corner. Click it and Melies will rewrite your prompt automatically - enriching it with more detail, better scene description, and motion cues. You can write a rough idea like "woman walks through rain" and the assistant will expand it into a detailed, model-friendly prompt. It is the fastest way to go from a simple concept to a production-quality prompt.
Image-to-Video Prompts
When starting from an image, your prompt should focus on what moves and how:
Good prompts:
The woman slowly turns her head toward the camera, hair moving gently in the wind
Camera slowly dollies forward toward the building entrance, people walking in the background
Rain begins to fall, the character looks up at the sky, water droplets on the lens
Less effective prompts:
A beautiful woman standing in the rain (too static - this describes an image, not motion)
Everything moves (too vague - the model needs specific direction)
Text-to-Video Prompts
When generating from text only, describe both the scene and the motion:
A detective walks through a neon-lit Tokyo alley at night, rain falling, camera following from behind, puddles reflecting neon signs
Close-up of a hand slowly opening an old leather journal, dust particles floating in golden afternoon light
Video Prompt Tips
- Start with the main action - What is the primary movement in the scene?
- Describe camera movement - Is the camera static, panning, dollying, or tracking?
- Keep it focused - One clear action per clip works better than multiple complex movements
- Include environmental motion - Wind, rain, background activity adds life
- Specify pace - "Slowly" vs "suddenly" vs "gradually" affects the animation speed
Step 3: Set Duration and Resolution
Duration
Duration varies by model:
- 3-15 seconds: Kling v3 models (most flexible)
- 4-8 seconds: Veo 3.1
- 2-12 seconds: Seedance v1 Pro
- 6 or 10 seconds: Hailuo 02
- ~5 seconds: WAN v2.2
- ~4.8 seconds: LTX 2 Pro
Tip: For narrative work, shorter clips (3-5 seconds) often look better than long ones. Each clip should capture one action or moment. You can combine multiple short clips on the timeline for longer sequences.
Resolution
- 4K: Veo 3.1 only
- HD/1080p: Kling v3 Pro, Kling v3 Standard, Kling O3 Standard, Seedance v1 Pro, LTX 2 Pro
- 768p: Hailuo 02 (maximum)
- 720p: WAN v2.2 (maximum)
For final production, aim for HD or higher. For tests and iteration, lower resolutions save credits and generate faster.
Step 4: Generate Your Video
Click Generate. Video generation takes longer than images - typically 30 seconds to a few minutes depending on the model and duration.
Once your video is ready, you can:
- Play it back to evaluate the motion and quality
- Generate again with the same settings for a different result
- Adjust the prompt if the motion was not what you expected
- Try a different model to compare results
Step 5: Extend Your Video
Need a longer clip? You can extend a generated video by generating a continuation. The model picks up where the previous clip left off and continues the motion.
This is useful for:
- Building longer sequences from a single starting image
- Letting a scene develop over more time
- Creating smooth, continuous motion beyond a single clip's duration limit
Step 6: Edit on the Timeline
Once you have multiple video clips, bring them to the Melies timeline editor to:
- Arrange clips in sequence to build scenes
- Trim and cut clips to remove unwanted frames
- Add transitions between clips
- Add audio - music, voice-over, sound effects with volume control
- Set the pace of your film through edit rhythm
Step 7: Export Your Film
When your edit is complete, export from the timeline:
- Formats: MP4 or WebM
- Resolutions: 480p, 720p, or 1080p
- Quality: Low (CRF 28), Medium (CRF 23), or High (CRF 18)
- FPS: 1-60 (default 25)
- Aspect ratios: 16:9, 3:2, 1:1, 2:3, or 9:16
Choose your export settings based on the destination - 1080p High for YouTube, 9:16 720p for TikTok/Instagram Reels, 16:9 1080p for general distribution.
The Complete Video Workflow
Here is the recommended workflow for creating professional AI video content:
- Start with images - Use the to create your scene
- Refine with variations - Get the perfect camera angle, lighting, and mood
How to Generate AI Image VariationsLearn how to generate AI image variations in Melies. Change camera angles, lighting, expressions, color grades, and more from a single base image. - Use AI actors for character consistencyacross scenes
AI Character Generator: Create Consistent AI CharactersUse the Melies AI character generator to create consistent characters across scenes. Browse 148 pre-built AI actors and generate images with the same character every time. - Generate video clips from your best images using image-to-video
- Iterate - Adjust prompts, try different models, regenerate until each clip is right
- Edit on the timeline - Arrange clips, add cuts, set pacing
- Add audio - Music, dialogue, sound effects
- Export in the right format for your platform
Model Comparison for Common Scenarios
Best for Film Quality
Veo 3.1 (400 credits) - 4K with native audio. Use for hero shots and final production.
Best Overall Value
Kling v3 Pro (100 credits) - HD, up to 15 seconds, multi-shot, voice ID. The best feature set at a reasonable price.
Best for Long Clips
Kling v3 Standard (60 credits) - Up to 15 seconds at HD for just 60 credits.
Best for Camera Control
LTX 2 Pro (50 credits) - Camera LoRA for precise dolly, jib, and static camera movements.
Best Budget Option
LTX 2 Pro (50 credits) or WAN v2.2 (60 credits) - Cheapest options for quick iteration.
What to Do Next
- Generate AI images- Create the perfect starting frames for your videos
How to Generate AI Images with MeliesLearn how to generate AI images step by step with Melies. Choose from 16 models, write effective prompts, pick the right aspect ratio, and create stunning visuals. - Compare video models- Deep-dive model comparison
Kling vs Veo vs Hailuo: Best AI Video Model ComparisonCompare Kling v3, Veo 3.1, Hailuo 02, WAN v2.2, Seedance, and LTX 2 Pro AI video models. Quality, speed, credits, duration, and features compared for 2026. - Create AI actors- Consistent characters across your video scenes
AI Character Generator: Create Consistent AI CharactersUse the Melies AI character generator to create consistent characters across scenes. Browse 148 pre-built AI actors and generate images with the same character every time. - Create a full AI film- The end-to-end filmmaking guide
AI Filmmaking: How to Create an AI Film from Idea to ExportLearn the complete AI filmmaking workflow with Melies. Generate ideas, cast AI actors, create storyboards, produce video clips, edit on a timeline, and export your finished AI film.
Visit the
to start creating, and check the for credit packages.
