How to Create Music Videos Using AI
CloneVoice | MusicCreator | TalkingPhotos | VideoExpress
Creating professional-looking music videos no longer requires a studio, camera crew, singers, or expensive equipment. With a handful of specialized AI tools, you can generate original songs, realistic singing characters, cinematic B-roll footage, and a polished final video, all from your computer.
In this step-by-step guide I share the exact workflow I use to create music videos. In the first section below, I cover the high-level process as shown in the overview video and then break it down into each step. For each step, you’ll also find a corresponding video tutorial.
The Overall Step-by-Step Workflow
Before we get into the nitty-gritty of each step of the process, in this section I walk you through the entire workflow. This will give you a good understanding of each of the components in creating engaging music videos. You end up with a full music video that mixes the singing performance with dynamic visual scenes that match the lyrics and mood. The complete process uses these core tools (with flexible alternatives noted). You can watch the complete overview below:
1. Generate an original song and lyrics (MusicCreator AI or CloneVoice AI).
2. Create a high-quality image of the singer (using an AI image tool such as Artistly, or ChatGPT/Grok).
3. Turn that still image + song into a lip-synced singing video (TalkingPhotos AI).
4. Use the song lyrics to generate scene-by-scene image and video prompts (Grok, ChatGPT, or others).
5. Convert those prompts into short cinematic B-roll clips (VideoExpress AI).
6. Composite everything in any video editor (CapCut, DaVinci Resolve, Movavi, etc.).
Step 1: Generate Your Song with AI
This foundational step produces the complete audio track (vocals + instrumentation) and the full lyrics that will drive every visual decision later. Using an AI music generation app like MusicCreator AI (prompt-based) or CloneVoice AI (genre- and theme-focused), you create a royalty-free song tailored to your desired style, mood, and length.
The lyrics become the creative blueprint for scene prompts, while the finished audio file is used for lip-syncing the singer. High-quality output here ensures the rest of the video feels cohesive and professional. You have two excellent options depending on your needs, but you can also use others like Suno.
Option A: MusicCreator AI (Prompt-Based Song Creation)
1. Log into MusicCreator AI and go to the Dashboard.
2. Select AI Music Generator.
3. Choose the Prompt mode (or switch to Own Lyrics if you already have words).
4. Write a detailed prompt that includes genre, mood, instruments, atmosphere, and vocal gender.
Example structure: “Upbeat indie pop with bright acoustic guitar and soft synth pads, warm female vocal, around 118 BPM, nostalgic but hopeful mood, clean modern production.”
5. Optionally select genre, mood, instruments, ambience, and vocal gender from the dropdowns.
6. Toggle the instrumental option if you want a track without vocals.
7. Click Generate Music. The system usually produces two versions in a few minutes.
8. Preview, download as MP3 or WAV, and copy the lyrics for later use.
Option B: CloneVoice AI (Genre + Lyrics Focused)
1. Log into CloneVoice and click Create Music. Use Version 2 (the latest).
2. Select a music genre (like gospel, pop, etc.).
3. Choose AI Generated lyrics (or paste custom lyrics in the required format).
4. Enter a short theme or title (e.g., “Jesus, my rock, my everything”).
5. Agree to the terms and click Create Lyrics. Review and edit the generated sections.
6. Enter a song title, agree to the terms again, and click Generate Music.
7. Wait for processing (often 10–15 minutes). Once complete, play, download, and copy the lyrics.
NOTE: Save both the audio file and the full lyrics, you will need them in the next steps.
Step 2: Create the Singer Image
You generate a realistic or stylized portrait of the performer that will serve as the consistent character throughout the project. Using tools like Artistly, Grok, or ChatGPT, you craft a detailed prompt describing appearance, clothing, lighting, expression, and framing (preferably a clear, forward-facing close-up or waist-up shot).
This single image becomes the reference for lip-sync generation and any character consistency features in later tools, so quality and consistency at this stage are critical. Use any strong image generator (I used Artistly, but you can use ChatGPT, Grok, Midjourney or any other AI image-generation app).
In the tutorial below, I walk you through some of the key features in Artistly AI. It’s a 20-minute video, however, you don’t need to watch the full video. The first 2.5 minutes will cover how to generate images. There are several ways to do so, but you can use the basic method as shown in the video.
1. Use Grok or ChatGPT for a detailed prompt describing the singer’s appearance, clothing, lighting, and style.
2. Generate a clear, forward-facing portrait (close-up or waist-up works best for lip-sync).
3. Download the final image. This becomes your consistent character reference.
Step 3: Create the Singing Photo / Video in TalkingPhotos AI
In TalkingPhotos AI (Singing V3), you upload the singer image and the generated song audio. The AI animates the still photo so the character sings with accurate lip movements, natural facial expressions, and subtle body motion. You choose aspect ratio (vertical or horizontal), style, and optional consistent-character settings. The result is a full-length performance video that forms the emotional core of the final music video. This is where the magic of lip-sync happens, however, the quality of the lip-sync will depend on the audio of the song. For best results, it’s recommended to isolate the vocals.
1. Log into TalkingPhotos AI and click Create Video.
2. Select Singing V3 (the improved version with better lip-sync and motion).
3. Choose a style (close-up high quality, fast, or with musical instrument).
4. Select aspect ratio: vertical (9:16) or horizontal (16:9).
5. Create a new character:
– Choose gender and type (Human, 3D Cartoon, or Animal).
– Enter an image prompt or upload your reference image.
– Generate and refine the face until you are happy.
6. Import the audio: either upload the MP3 or pull it directly from CloneVoice (optional).
7. Name the project and click Render Video.
8. Wait for processing (can take 10–20 minutes depending on length). Download the finished MP4.
The result is a full-length singing performance with natural lip movements and subtle body motion.
Step 4: Create Cinematic B-Roll Clips in VideoExpress AI
You feed the complete lyrics into an LLM such as Grok or ChatGPT and request a structured shot list. The AI breaks the song into logical scenes and writes detailed image prompts (what the frame looks like) plus corresponding video/action prompts (camera movement, lighting changes, motion, atmosphere). These prompts ensure the B-roll visuals directly match the story, emotion, and timing of the lyrics rather than feeling random or disconnected.
Copy the full song lyrics into Grok (or ChatGPT) and ask it to:
1. Break the song into logical scenes.
2. Write detailed image prompts for each scene.
3. Write corresponding video/action prompts that describe camera movement, lighting, and motion.
This gives you a ready-made shot list that matches the emotional arc of the lyrics.
In VideoExpress AI, you take each image and video prompt pair and generate short (typically 5–10 second) cinematic clips. You create the still image first (often with photorealistic cinematic styling and character consistency), then animate it with the action prompt, camera motion, and extended duration settings. Repeating this for every scene produces a library of dynamic B-roll footage that can be interleaved with the singing performance.
1. Log into VideoExpress AI and click Create with AI → Create Video from Prompt.
2. Choose the same aspect ratio you used for the singing video.
3. Paste an image prompt. Select a style such as “Photorealistic Cinematic.”
4. Enable Use Consistent Character and upload your singer reference image.
5. Generate the still image(s). Save the best one to your media library.
6. Paste the matching video prompt into the action box.
7. Enable Advanced Mode, extend the clip length (up to 10 seconds works well), and generate the video.
8. Repeat for every scene. All finished clips appear in your Media Library under My AI Videos.
Step 5: Composite the Final Music Video
Once you have all the above components ready, you can use a video editor to composite the final music video. You assemble the final music video by placing the full singing performance on the primary track and splitting it at natural section breaks.
The cinematic B-roll clips are layered on upper tracks and timed to the corresponding lyric sections. Where B-roll plays, it covers the singer; where it is absent, the performance remains visible.
Final touches such as color grading, transitions, and audio balancing complete a polished, professional-looking music video ready for YouTube, Shorts, Reels, or other platforms.
You can watch the workflow video tutorial as shown in introduction section (just before step 1 above). Towards the end of the video, I show you how I used Movavi video editor to composite the final music video. Listed below are the overall steps involved:
1. Place the full singing video from TalkingPhotos on the bottom video track.
2. Split the singing clip at natural lyric or musical section points. This makes it easy to insert B-roll.
3. Drop the VideoExpress B-roll clips onto the upper track(s), timed to the corresponding lyric sections.
4. Where B-roll is active, the singing performance is covered; where there is no B-roll, the singer remains visible.
5. Add any final touches like color grading, subtle transitions, text overlays, or volume adjustments.
6. Export in the desired resolution and format.
The result feels like a professionally directed music video even though almost every element was generated by AI.
Tips for Best Results
1. Keep the singer image consistent across tools by using the same reference photo.
2. Match aspect ratios (vertical or horizontal) throughout the project for consistency.
3. Generate more B-roll clips than you think you need; you can always cut them later.
4. Use the lyrics as the creative backbone; they drive both the performance and the visual storytelling.
5. All the tools mentioned produce royalty-free content suitable for monetization when used according to terms.
Anchor in the Deep Music Video
Below is the music video that I created, as shown in this step-by-step guide. It’s on my How To Tutorials Facebook page. To see it in full screen, once you click the Play button, make sure to click the “Full Screen” (double arrow) button in the bottom-right of the player.
AFFILIATE DISCLAIMER: If you use the links on this page to purchase any of the mentioned products, I may earn a commission as an affiliate marketer. This recommendation and review is based on my firsthand experience using the mentioned products myself.
