Deep Down In My Heart AI

What is that deep down in my heart AI crying meme everyone is posting on TikTok lately?

It is a two-character confession scene followed by an exaggerated crying reaction. The background reference is Rocky Lockridge’s emotional family moment in A&E’s Intervention, later known as Best Cry Ever. This deep down in my heart meme generator casts your two subjects into a 30-second interview and retains the template’s original dialogue and crying soundtrack. A depth reference guides the camera cuts, gestures and pauses, while your photos supply the identities. The phrase “deep down in your heart” can also describe a sincere feeling; here the tool refers specifically to this crying-video format.

How do I make the deep down in my heart AI video using two pictures?

Upload one clear adult or pet photo for each role, choose a resolution, then check the displayed credit total. This deep down in my heart video uses a fixed 30-second vertical template. Photo 1 supplies the crier; Photo 2 supplies the confessor. The workflow first prepares a storyboard from both photos, checks the reference assets, and asks Seedance 2.5 to follow the depth performance. It then adds the original soundtrack to the finished picture. No editing or audio upload is needed. Watch the complete result with sound before sharing: identity, mouth movement and hand or paw shapes can still vary.

1. 01

   Choose who cries

   Add one clear portrait in Photo 1. Keep the whole face visible, with neutral lighting and no hand covering the mouth.

2. 02

   Choose who confesses

   Add the confessor in Photo 2. Confirm permission, choose 480p, 720p or 1080p, and check the total for the complete 30-second workflow.

3. 03

   Generate and review

   We prepare the storyboard, generate the interview and add the original sound. Review both identities and the crying reaction, then download the MP4.

Which photo am I supposed to upload first for the deep down in my heart template?

Upload the crier first and the confessor second. These roles stay attached to the two image positions through the storyboard and video requests. Choose photos with one visible adult or pet, a clear mouth or muzzle and similar camera height. The template follows the interview’s alternating close-ups, so role order matters more than left or right. If you only have a joint photo, crop it into two separate portraits before uploading; each crop must still be at least 300 pixels on both sides. Cropping cannot recover missing facial detail. Replace the appropriate photo before generating if the roles are reversed.

How much does it cost to generate one deep down in my heart AI video?

One complete 30-second video costs 321 credits at 480p, 711 at 720p or 1,737 at 1080p. Seedance 2.5 is the only model on this page; 480p is the default. The single charge includes storyboard preparation, video generation and original-soundtrack assembly. The dollar equivalents allocate credits from Studio monthly at $49.90 for 1,800 credits; they are not checkout prices. The eligible 5-credit gift and $5 / 24-credit Trial are each insufficient alone. From zero, the $29.90 / 700-credit Boost pack covers 480p. Creator monthly includes 800 credits, enough for 720p; Studio monthly covers 1080p.

Equivalent cost uses $49.90 / 1,800 credits. Purchase amounts, remaining balance and account eligibility are separate. Existing plan prices are unchanged.

Is Starrd or MakeViral better for making the deep down in my heart AI crying meme?

Choose by the advertised output and the amount you must buy, rather than comparing credit counts across websites. Starrd lists a 30-second version; MakeViral lists a 15-second version; Jump on the Trend lists a 10-second version with a free still preview. TryH3 offers a 30-second depth-guided interview with original sound and selectable resolution. A free still does not mean a free finished video. The comparison records public page information checked on October 28, 2026. Unstated allowances and watermark terms are marked below instead of inferred. These are specifications, not results from a paid side-by-side generation test.

Why does my face look completely different halfway through the Best Cry Ever AI video, and how can I fix it?

Large expression changes, hands over the mouth and cuts between close-ups can make a model redraw facial details. Start with evenly lit portraits, a visible jawline and no strong beauty filter. Give each role a separate image instead of using the same group photo twice. If a face changes after a cut, try a clearer portrait before spending credits on another attempt. Higher resolution increases output size but does not guarantee identity accuracy or correct lip sync. A role-specific prompt helps express the sequence; it cannot guarantee that every generated frame follows it.

What this tool cannot do

This tool cannot guarantee frame-exact cuts, perfect identity or exact lip sync, even with a timed depth reference and the original audio. It does not repair an existing clip or generate a different script. Small or obstructed faces can drift during the crying close-ups. Pets are supported in the prompt, but paws can become hand-like and speech can distort a muzzle. Both subjects may still move their mouths, a cry may look like a smile, and teeth may deform. The original soundtrack is fixed; optional scene notes affect visual direction only. Review the result before posting it as a clearly fictional video.

Can you give me a copy-paste AI video prompt for the deep down in my heart Best Cry Ever meme?

These are the two base prompts used by this page. The first prepares a storyboard from your ordered photos. The video prompt uses Photo 1 as the crier, Photo 2 as the confessor, the storyboard as Image 3 and a 30-second depth performance as Video 1. It specifies eight shots and mouth movements aligned with the reference sound. Your optional visual note is appended after the base instructions. The final original soundtrack is added separately. Copying the video prompt alone into a text-only generator will not reproduce this workflow; the three image references and the depth video are also required.

1. Storyboard prompt

Create one vertical 9:16 photorealistic storyboard sheet in a clean 2-column by 3-row grid, six equal panels with narrow plain gutters and no text or numbers. This sheet plans an obviously fictional confession-and-crying interview parody. Use the first reference photo only for THE CRIER and the second reference photo only for THE CONFESSOR. Preserve each subject's identity, hair or fur, clothing and species consistently in every panel. Never blend or swap them. Adults and pets are supported; animals keep natural animal anatomy, paws, ears and muzzles.

Stage both subjects together on a sofa in the location and lighting of the first photo. Maintain one coherent room, sofa, wardrobe and screen direction across all six panels. Do not copy any people, outfits or text from a television clip.

Reading left to right, top to bottom: (1) the opening interview two-shot, CRIER facing the camera with a guarded expression, CONFESSOR nearer the camera seen partly from behind; (2) close-up of CONFESSOR with wet eyes, a hesitant half-smile and lips parting to confess; (3) close-up of CRIER holding a hand or natural paw over the mouth, eyes welling up; (4) close-up of CRIER in a sustained open-mouth wail, brows squeezed up, tear-wet cheeks, a runny nose and shoulders shaking; (5) close-up of CONFESSOR half laughing through tears; (6) close-up of CRIER resting the chin in one hand or paw, still sobbing.

Make the crying expressive but keep the subject recognizable. Realistic eyes, teeth and hands; no extra limbs or subjects. Pets have real paws, never human fingers. No subtitles, logos, watermarks or speech bubbles.

2. Video prompt

Create a 30-second vertical 9:16 photorealistic confession-and-crying interview parody. @图片1 is THE CRIER, @图片2 is THE CONFESSOR, and @图片3 is the prepared six-panel storyboard for their identities, expressions and room. @视频1 is a grayscale depth performance with the original dialogue and crying audio. Use its movement, framing, cuts, speaking windows and rhythm; render a full-color scene with the two uploaded subjects. Do not render the grid, grayscale depth surfaces or original interview subjects.

Preserve the identity, hair or fur, clothes and species from each subject photo. Keep the same sofa, location and lighting shown in the storyboard. Never exchange the two roles. Pets remain animals with natural paws, muzzles and proportions, even in expressive close-ups. No added people.

Match these eight shots and exact cut times: 0.0–1.6s opening two-shot with CRIER facing camera and CONFESSOR partly seen from behind; 1.6–6.8s CONFESSOR close-up, emotional confession; 6.8–9.1s CRIER close-up, silent with hand or paw over mouth; 9.1–13.4s CONFESSOR close-up, finishes the confession and pauses; 13.4–25.5s CRIER breaks into sustained sobs and an open-mouth wail, strongest near 19s, then chin in hand or paw from 21s; 25.5–26.6s tighter CRIER close-up, still sobbing; 26.6–29.4s CONFESSOR laughs through tears; 29.4–30.0s final CRIER close-up, sobbing into one hand or paw.

Follow the audio in @视频1 for all mouth movements. Only CONFESSOR mouths the opening confession and the later words "I still love you" in their original speaking windows. CRIER remains silent until the breakdown and then opens and closes the mouth with the original wails and breaths. Keep pauses and cuts fixed so the original soundtrack can be laid back onto the completed video without retiming. Do not invent dialogue or move a line to the other subject.

Crying must be visibly intense: wet cheeks, crumpled brows, tears, a wet nose, trembling shoulders and sustained open-mouth sobbing, while maintaining facial identity. Natural teeth, eyes and five fingers on human hands; real paws on animals. No captions, text, logos, channel badges or watermarks. Produce one continuous full-frame video, never a split screen or storyboard.

This requests consistent identity and lip sync; it does not guarantee either.

Can I make a deep down in my heart AI video completely free?

The standard signup gift alone does not cover this workflow: the minimum is 321 credits, while eligible new accounts receive up to 5. You can read the guide and watch the example without generating anything. The prepared storyboard is part of the paid workflow, not a free preview. Existing credits may cover an attempt, so check your balance before purchasing. The questions below cover photo order, pets, original sound and output limits. Higher resolution does not guarantee better identity or mouth movement. A completed result with visual defects still uses credits; a failed workflow returns the whole charge.