Turn your Suno song into a music video with a consistent AI singer who lip-syncs

1. Workflow

1. Map the song before you generate anything: section timestamps, which sections get lip-synced performance shots (usually the chorus), a one-paragraph singer description, your locations, and a numbered shot list with each shot 30 seconds or shorter.

   Tool

   Any chat (Claude, ChatGPT or Grok)

   Why

   Higgsfield's music video guide says to break the track into sections first because each plays a different role: intro sets the world, verses carry story or atmosphere, the chorus gets the strongest performance shots, the bridge changes the look (published Oct 1, 2026, checked Oct 21, 2026). Lip-syncing only some shots is also what keeps the budget down (see costs).

   Cost

   $0 on free plans.

   Copy prompt Paste into any chatbot, with your lyrics and the section start times from the Suno player.

   Preview prompt: Plan my music video (section map and shot list)

   ROLE You are a music video director and editor who plans AI-generated music videos around a finished song. GOAL Help me turn my finished song into a shot plan for a music video with one consistent AI singer, so every shot has a job and a timestamp before I generate anything. WHAT THE NEXT STEP NEEDS For Step 2 and Step 3: the exact timestamps of each song section, which sections get lip-synced performance shots, and a one-paragraph description of the singer. For Step 4: a numbered shot list. HOW TO WORK WITH ME Do not write the deliverable yet. Interview me first. - Ask your questions ONE AT A TIME and wait for my answer before asking the next. (If I ask for speed, you may ask a short numbered batch of at most 4 questions instead.) - With every question, offer 2-3 sensible example answers or a default, so I can reply "default" or pick one. - If an answer is vague, ask one follow-up. If I don't know, choose a sensible default and tell me what you chose. - Do not generate the deliverable until I have answered every question. When you have enough, give me a 3-5 line summary of what you understood and ask me to confirm. - Do not invent facts about my song, my lyrics or the tools. If you are unsure, say so. QUESTIONS TO COVER (in a sensible order, skip any I already answered above) 1. Paste the lyrics with section labels (intro, verse, pre-chorus, chorus, bridge, outro) and the start time of each section from the song player. (Example: "Verse 1 0:12, Chorus 0:41".) 2. Song title, genre, tempo feel and mood in a few words. (Example: "late-night synth-pop, 104 BPM, bittersweet".) 3. Who is the singer on screen? An invented AI character, or me (my own face)? Describe age range, look, styling and one signature item. (Default: an invented character with one signature item, like a red jacket, that appears in every shot.) 4. What is the world of the video: one location, a few locations, or a simple story? (Examples: "rooftop at night, car interior, empty diner"; "a breakup told in three scenes".) 5. Which cut are you making first: a 30-60 second cut for Shorts/Reels/TikTok around the chorus, or the full song? (Default: the chorus cut first, full song later.) 6. Aspect ratio: 16:9 for YouTube, 9:16 for Shorts/Reels/TikTok, or both? (Default: 9:16 for the short cut, 16:9 for the full video.) 7. Any visual references: films, eras, colors, camera style? (Examples: "90s camcorder", "35mm film, warm grain", "clean modern".) THE DELIVERABLE (only after I confirm) 1. SECTION MAP: one line per section with start-end timestamps, its visual role (intro = establish the world, verse = story or atmosphere, chorus = strongest performance shots, bridge = a visual change, outro = resolve), and whether it gets a lip-synced performance shot (YES/NO). 2. SINGER BIBLE: one paragraph I can paste into Higgsfield (look, face, hair, wardrobe, signature item, what never changes). 3. LOCATION LIST: each location in one line (setting, time of day, light, color). 4. SHOT LIST: numbered shots, each with timestamp in-out (keep each shot 30 seconds or shorter), type (PERFORMANCE / STORY / ATMOSPHERE), location, what happens, camera move, and which lyric line is sung if it is a performance shot. 5. A count of how many seconds of performance shots and how many seconds of other shots the plan needs. Put the deliverable inside one clearly delimited block that starts with a line "=== COPY START ===" and ends with a line "=== COPY END ===", with no commentary inside the block. Put any notes outside it. AFTER THE DELIVERABLE Ask me: "Anything to tweak?" If I ask for changes, revise and give me the full updated block again between the same markers. Repeat until I say it is good. FINISH When I'm happy, end with one line telling me what to do next: save this plan, then in Step 2 download the song and its lead-vocal stem from Suno and cut the vocal clips for every PERFORMANCE shot.

2. In Suno, download the song (WAV on Pro), then open the song's More Actions (...) menu > Get Stems > Split from Mix and pick Lead Vocals to get the vocal on its own. In any editor, cut a vocal clip for each PERFORMANCE shot using the timestamps from Step 1.

   Tool

   Suno (Pro or Premier)

   Why

   Suno's help center says Get Stems is for Pro and Premier, and Split from Mix extracts one vocal or instrument plus a complement track at 10 credits per stem (article edited Sep 9, 2026, checked Oct 21, 2026). Suno's pricing page lists WAV downloads, stem separation and commercial use rights on Pro and Premier, and none of the three on Free (checked Oct 21, 2026). Our tip, not a vendor claim: a clean vocal clip gives the lip-sync step less music to confuse the mouth with; keep the full mix for the final edit.

   Cost

   Inside Suno Pro: 10 credits for the vocal stem, of 2,500 monthly credits.

3. Create your singer once and lock the look: build the character in Higgsfield AI Influencer (Character Builder), or use Add your face if the singer is you, generate 20+ clear images of the same face, then train a Soul ID from them. Make one reference frame per location with Soul Cinema.

   Tool

   Higgsfield (AI Influencer, Soul ID, Soul Cinema)

   Why

   Higgsfield's changelog says AI Influencer builds a character from scratch, from a reference, or from your own photo, with 19 Character Builder categories (Oct 2, 2026). Its music video guide says Soul ID trains a reusable identity from 20+ reference photos so the face holds across later generations, and Soul Cinema makes the still frames that set locations, lighting and style (checked Oct 21, 2026). Only use a real person's face or voice with their permission; Higgsfield's guide lists permission for likeness and voice as a prerequisite.

   Cost

   Higgsfield's guide example: Soul ID setup 25 credits; a Soul Cinema reference image 0.125 credits.

4. Turn the shot list into one short video prompt per shot, with the same camera look and color palette throughout, and the singer called by the name you saved.

   Tool

   Any chat (Claude, ChatGPT or Grok)

   Why

   Cinema Studio 4.0 exposes genre, camera and lens, tempo, emotion, color palette and era controls (Higgsfield changelog, Aug 12, 2026; music video guide, Oct 1, 2026). Writing every prompt in one pass keeps those settings and the singer description identical from shot to shot, which is the cheapest way to avoid drift.

   Cost

   $0 on free plans.

   Copy prompt Paste into the same chat as Step 1, or a new one with your Step 1 plan.

   Preview prompt: Write my shot prompts

   ROLE You are a prompt writer for AI video generation. You turn a music video shot list into short, concrete video prompts. GOAL Write one ready-to-paste video prompt per shot from my shot list, so the singer, locations and look stay the same across every generation. WHAT THE NEXT STEP NEEDS For Step 5: a numbered prompt for every shot, with the length in seconds and the settings to pick (genre, camera/lens, tempo, emotion, color palette, era), in the same order as the shot list. HOW TO WORK WITH ME Do not write the deliverable yet. Interview me first. - Ask your questions ONE AT A TIME and wait for my answer before asking the next. (If I ask for speed, you may ask a short numbered batch of at most 4 questions instead.) - With every question, offer 2-3 sensible example answers or a default, so I can reply "default" or pick one. - If an answer is vague, ask one follow-up. If I don't know, choose a sensible default and tell me what you chose. - Do not generate the deliverable until I have answered every question. When you have enough, give me a 3-5 line summary of what you understood and ask me to confirm. - Do not invent facts about my song, my lyrics or the tools. If you are unsure, say so. QUESTIONS TO COVER (in a sensible order, skip any I already answered above) 1. Paste the SECTION MAP, SINGER BIBLE, LOCATION LIST and SHOT LIST from Step 1. (Required.) 2. What did you name your singer's saved identity or Element in Higgsfield? (Example: "@Mara". Default: "@Singer".) 3. Which settings do you want to keep fixed for the whole video? (Examples: "35mm film camera, warm palette, 1990s era"; default: one camera look and one color palette for every shot.) 4. Should performance shots look straight into the camera, or past it? (Default: chorus shots to camera, verse shots past it.) 5. Anything that must never appear? (Examples: text on screen, logos, other people's faces, real brands.) THE DELIVERABLE (only after I confirm) For each shot, in order: - SHOT number, timestamp in-out and length in seconds (30 seconds or shorter). - SETTINGS: genre, camera/lens, tempo, emotion, color palette, era (only the ones that change from the fixed settings). - PROMPT: 2-4 sentences in plain English: who (the singer by their saved name), where, what they do, camera move, light. For PERFORMANCE shots, say the singer is singing with clear mouth movement, face visible and well lit, facing the camera or three-quarter angle, so lip-sync has something to work with. - AVOID: one short line of what must not appear. After the last shot, list the reference images I should attach (singer identity, each location frame). Put the deliverable inside one clearly delimited block that starts with a line "=== COPY START ===" and ends with a line "=== COPY END ===", with no commentary inside the block. Put any notes outside it. AFTER THE DELIVERABLE Ask me: "Anything to tweak?" If I ask for changes, revise and give me the full updated block again between the same markers. Repeat until I say it is good. FINISH When I'm happy, end with one line telling me what to do next: generate each shot in Step 5, starting with the PERFORMANCE shots, and keep a note of which take you kept for each shot number.

5. Generate the shots in Cinema Studio 4.0 with your Soul ID and location frames attached: performance shots first (singer facing camera, face lit, clearly singing), then story and atmosphere shots. Keep the best take per shot number.

   Tool

   Higgsfield Cinema Studio 4.0

   Why

   Higgsfield says Cinema Studio 4.0 runs on Seedance 2.5, generates up to 30 seconds per clip and takes up to 50 references per generation (music video guide, Oct 1, 2026; changelog, Aug 12, 2026). Forward / Backward Extend can continue or lead into a clip you like instead of starting over (changelog, Aug 12, 2026).

   Cost

   Higgsfield's guide example: 240 credits for one 20-second shot (about 12 credits a second, our arithmetic). Retries add to this.

6. Run each performance shot through Lipsync Studio with its vocal clip from Step 2, check the mouth against the words, and redo only the shots that drift.

   Tool

   Higgsfield Lipsync Studio

   Why

   Higgsfield's guide says Lipsync Studio matches the singer's mouth to the vocal track across 18 languages and seven models, and that this is the step that makes a generated performance shot match the song being sung (Oct 1, 2026). Its lip-sync guide adds that clips under about 30 seconds, cut together, avoid repetitive gestures (Jul 7, 2026).

   Cost

   Higgsfield's guide example: 67 credits for a 20-second performance shot.

7. Edit and release: put the full Suno song on the timeline first, cut each shot to its timestamp, mute the generated clips' own audio, export 16:9 and a 9:16 chorus cut, then use the prompt to check the edit and write the upload text. On YouTube, set Attributes > AI use to Yes.

   Tool

   Any video editor (e.g. CapCut, DaVinci Resolve; not one of our tools) + any chat

   Why

   Higgsfield's guide says to place the finished song on the timeline first, cut shots to the planned timestamps and mute generated clip audio where it conflicts with the track (Oct 1, 2026). YouTube's help center says creators must disclose AI-generated content that creates music that's the main focus of the video, using the AI use setting in YouTube Studio (checked Oct 21, 2026).

   Cost

   $0 for the chat; editors have free tiers (not priced here).

   Copy prompt Paste into any chatbot after your first full export.

   Preview prompt: Check my edit and write the upload

   ROLE You are a release checker for AI music videos. You review an edit before upload and write the upload copy. GOAL Catch the problems viewers notice first (face changes, mouth out of sync, continuity jumps) and prepare an honest title, description and AI disclosure for the upload. WHAT THE NEXT STEP NEEDS A short fix list for the edit, and paste-ready upload text: title, description, tags and the AI disclosure line, for YouTube (and a short caption for Shorts/Reels/TikTok if I want one). HOW TO WORK WITH ME Do not write the deliverable yet. Interview me first. - Ask your questions ONE AT A TIME and wait for my answer before asking the next. (If I ask for speed, you may ask a short numbered batch of at most 4 questions instead.) - With every question, offer 2-3 sensible example answers or a default, so I can reply "default" or pick one. - If an answer is vague, ask one follow-up. If I don't know, choose a sensible default and tell me what you chose. - Do not generate the deliverable until I have answered every question. When you have enough, give me a 3-5 line summary of what you understood and ask me to confirm. - Do not invent facts about my song, my lyrics or the tools. If you are unsure, say so. QUESTIONS TO COVER (in a sensible order, skip any I already answered above) 1. Paste your shot list and tell me, shot by shot, anything you noticed: face looks different, mouth off the vocal, outfit or signature item changes, jump in lighting. (Example: "shot 6 the jacket turns blue; shot 9 lips late by a beat".) 2. Where will you post: YouTube full video, Shorts/Reels/TikTok cut, or both? (Default: both.) 3. Artist name, song title, release date and where the song streams, if it does. (Leave blank if not released yet.) 4. Is the singer an invented AI character or your own face? Is the music AI-generated? (Default: invented character, AI-generated music.) 5. Tone for the copy: plain, hype, or behind-the-scenes? (Default: plain with a short behind-the-scenes line.) THE DELIVERABLE (only after I confirm) 1. FIX LIST: for each problem shot, the cause to check (identity reference, prompt, audio clip timing) and the cheapest fix (re-cut, re-sync, regenerate), ordered by how visible it is. 2. FINAL CHECK: a 10-item yes/no checklist: likeness, lip sync on every performance shot, continuity of the signature item, shot timing on the beat, transitions, generated clip audio muted under the song, loudness, safe text margins for 9:16, end card, file exported at the right size. 3. UPLOAD COPY: title (under 70 characters), description (3-5 short lines plus credits: song made with Suno, video made with Higgsfield), 5 tags, and one plain sentence stating that the music and video were made with AI tools. 4. A reminder line: "YouTube Studio > upload > Attributes > AI use: Yes". Never claim a real person sang or appears in the video unless I told you so. Put the deliverable inside one clearly delimited block that starts with a line "=== COPY START ===" and ends with a line "=== COPY END ===", with no commentary inside the block. Put any notes outside it. AFTER THE DELIVERABLE Ask me: "Anything to tweak?" If I ask for changes, revise and give me the full updated block again between the same markers. Repeat until I say it is good. FINISH When I'm happy, end with one line telling me what to do next: fix the shots on the list, export, then upload with the AI use setting turned on.

2. Recommended tool chain

1. Any chat (Claude, ChatGPT or Grok)
2. Suno (Pro or Premier)
3. Higgsfield (AI Influencer, Soul ID, Soul Cinema)
4. Any chat (Claude, ChatGPT or Grok)
5. Higgsfield Cinema Studio 4.0
6. Higgsfield Lipsync Studio
7. Any video editor (e.g. CapCut, DaVinci Resolve; not one of our tools) + any chat

3. Total cost

Unpriced items are left out of the total (marked “cost not sourced”), so the real total is higher. All totals are our arithmetic from sourced unit prices.

4. Current promos

• Suno annual billing: Pro saves $24 a year

  Suno's pricing page lists Pro at $8/month billed yearly and says yearly billing saves $24 (Premier: $24/month, saves $72). Standing discount, not a time-limited sale.

  Applies to: Suno Pro in Step 2 (our routes use monthly billing for a one-month project)

  Start date from the first readable archive snapshot of the pricing page; the discount may be older.

  Source: Suno: Pricing, checked Oct 21, 2026

• Higgsfield annual billing: Plus $47/month instead of $59

  Higgsfield's pricing page listed Plus at $47/month billed annually versus $59 monthly (saves $144/yr) and Ultra at $99/month annually versus $129 monthly (saves $360/yr) when we read it on Oct 15, 2026.

  Applies to: Higgsfield steps 3, 5 and 6 if you'll keep making videos

  Standing annual price; start date not checked.

  Source: Higgsfield: Pricing, checked Oct 15, 2026

5. Learn it

Last verified Oct 21, 2026. Items older than 30 days are marked.

6. Caveats and assumptions

• Credit costs in this playbook are the examples in Higgsfield's Oct 1, 2026 guide, not a price list; Higgsfield shows the exact cost before each generation, and retries add up quickly.
• Higgsfield's pricing page didn't render for us today, so plan prices are from our Oct 15, 2026 check.
• Lip-sync can still miss fast or slurred lines. Shorter performance clips are easier to redo than long ones.
• Commercial use: Suno lists commercial use rights on Pro and Premier and none on Free, and says to see its terms of service for limits (checked Oct 21, 2026). Check Higgsfield's terms for your plan before monetizing.
• Don't put a real artist's face or voice in a video without their permission. Higgsfield's own guide lists permission for the music, likeness and voice as a prerequisite.

Last verified Oct 21, 2026.