Trouble Maker AI Dance From Two Separate Photos: Tested Results, Framing Limits & Credits

Yes—on the Trouble Maker AI Dance template at HotDance AI (hotdanceai.com), you can upload two completely separate individual portraits (JPG, PNG, WEBP, or HEIC) directly into the Photo of Dancer 1 and Photo of Dancer 2 slots without combining them into one image. All two-photo runs automatically route to Seedance 2.5 (12s watermark-free MP4 with synced audio) at 17 credits (Standard 480p), 37 credits (HD 720p), or 93 credits (Full HD 1080p) after an instant 0-credit GPT Image 2.6 still-frame preview. In verified production smoke tests (2026-11-09 and 2026-11-10), two separate standalone portraits successfully replaced both dancers, and safety-blocked runs refunded 100% of deducted credits automatically—though pairing one tight headshot specifically against one waist-up photo through rapid crossings has not yet been controlled-benchmarked.

Below are the exact slot rules, verified production findings vs. unbenchmarked framing edge cases, and clear credit/refund mechanics before you render.

Trouble Maker AI Dance Template Specs & Credit Matrix

Because two-person partner turns and crossings demand higher-capacity identity binding, HotDance AI routes every 2-photo duet through Seedance 2.5 rather than lower solo tiers.

Setting / Rule	Verified Behavior on Trouble Maker AI Dance
Input Slots & Formats	2 separate slots: Photo of Dancer 1 and Photo of Dancer 2 (JPG, PNG, WEBP, HEIC). One consenting 18+ adult per slot; photos do not need to be taken in the same room or day.
Consent & Safety Rules	Explicit consent from both adults (18+) required; minors, public figures, and unauthorized celebrity likenesses are prohibited.
Pre-Pay Still Preview	0 credits, no sign-up required — OpenAI GPT Image 2.6 places both faces into a key stage frame before purchase (capped at 3 per template, 15 total per visitor, 15 per IP/day). Not a playable video preview.
Seedance 2.5 Output Specs	12-second vertical MP4 (720×1280 verified at HD 720p), 100% watermark-free, pre-muxed synced .m4a AAC soundtrack (with optional mute toggle).
Exact Credit Rates	17 credits Standard 480p (default) · 37 credits HD 720p · 93 credits Full HD 1080p.
One-Time Pack Math	Packs from $9.90 (10 cr), $17.99 Trio Pack (30 cr — covers a 17-cr Standard 480p duet with 13 credits remaining), or $29.00 Squad Pack (50 cr — covers a 37-cr HD 720p duet with 13 credits remaining). Unused credits never expire.

What Production Tests Verified—and What Remains Unbenchmarked

Before choosing source photos, separate what has actually been verified on live production workers from edge cases we have not run controlled A/B tests on:

• Verified: Two separate standalone portraits replace both dancers (2026-11-09 & 2026-11-10 smoke tests). Running /trouble-maker-ai-dance with two independent individual portraits (image-source.jpg + ref.jpg) generated a 720×1280 vertical MP4 in which both performers took on the uploaded subjects' faces and appearance alongside synced .m4a audio.
• Verified contrast: Why 2 slots beat 1 slot on multi-person scenes. In a companion smoke test on APT AI—where the underlying video features two performers on screen while the form accepts only 1 photo—the generator replaced only the lead performer and left the second person in the scene unchanged. Always pick a dedicated two-slot template in our Couple & Duo Dance hub whenever you need two people replaced.
• Verified: 100% automatic refund on upstream content-safety blocks. During 2026-11-09 production testing, one Trouble Maker render triggered an upstream provider safety filter (OutputVideoSensitiveContentDetected) after submission; HotDance AI automatically canceled the job and returned 100% of deducted credits immediately (0 net credits charged). Because duets involve tight close-ups and dancing attire, intermittent upstream moderation triggers can happen even on valid inputs—making atomic refunds essential.
• Unbenchmarked edge case: Pairing 1 tight headshot with 1 waist-up photo through rapid turns & occlusion. HotDance AI has not conducted a controlled benchmark specifically comparing mismatched framing (for example, a neck-up headshot in Dancer 1 paired with a waist-up shot in Dancer 2) during close shoulder crossings and fast turns. Because the model must invent unseen clothing and torso proportions for the tight headshot while tracking two crossing silhouettes, mismatched crop scales carry higher risk of proportion shift, sleeve bleed, or momentary identity wobble—match both portraits at roughly waist-up or three-quarter framing whenever possible.

How to Pick Two Portraits That Hold Identity Best

Because the generator tracks two independent identities inside one choreography path, matching framing upfront prevents most blend artifacts:

1. Match framing scale across both slots: Pair two waist-up or full-body portraits (rather than one passport close-up and one wide mirror shot) so both characters share similar visual detail density and wardrobe cues.
2. Use single-subject portraits: Make sure each uploaded image shows only one person clearly facing forward—no background friends, sunglasses, microphones, or hands covering the jaw.
3. Keep lighting clean and comparable: Sharp, evenly lit photos let Seedance 2.5 hold eye spacing, jawline, and hairstyles across spins without dark shadow smearing.
4. Verify slot assignment in the free still preview: Upload both photos on Trouble Maker AI Dance and inspect the 0-credit GPT Image 2.6 still-frame preview first so you confirm how each face maps onto Dancer 1 and Dancer 2 before confirming the paid render.

Stage Selection, Credit Deduction & Refund Guarantees

Below the two photo slots, select an available stage preset (such as K-Pop Arena Stage) and choose your output resolution (17 cr 480p, 37 cr 720p, or 93 cr 1080p).

• When credits deduct: Photo uploads and the GPT Image 2.6 still preview cost 0 credits. Credits deduct atomically when you confirm and submit the paid MP4 job (after signing in or completing passwordless Stripe / Apple Pay / Google Pay guest checkout, which returns directly to the template).
• When credits refund automatically: If the task fails, times out, is canceled before provider dispatch, or triggers an upstream moderation block, 100% of deducted credits refund automatically (src/modules/ai-tasks/service.ts).
• Completed bad renders & retention: Under Refund Policy, one-time packs are final sale in cash (except duplicate billing), while confirmed broken swaps can be reviewed by support@hotdanceai.com for credit restoration. Uploads and outputs are deleted automatically within 30 days.

Template Workflow vs. Custom Reference Videos

Unlike DIY motion-transfer tools where you hunt for a clean driving MP4 yourself, HotDance AI does not require or accept custom reference-video uploads on /trouble-maker-ai-dance or /body-swap. Use dedicated two-slot duet presets like Trouble Maker AI Dance, Laugh Now Cry Later AI, or Hotel Lobby AI whenever two consenting portraits need to hit beat-synced choreography together.

Audio Playback vs. Publishing License

Play your downloaded MP4 locally to verify sound sync, and keep playback distinct from music licensing: an included template soundtrack does not convey commercial synchronization rights for recognizable K-pop or chart music. For monetized channels or branded posts, toggle sound off before exporting or swap the audio inside TikTok's or Reels' approved music library.

Open the Trouble Maker AI Dance page with two matched, consenting adult portraits, verify your 0-credit character still preview, and confirm the live Seedance 2.5 quote (17–93 credits) before rendering.