The "somewhere deep down in my heart" meme comes from a 2010 TV interview where one man says those words to another, who then bursts into loud tears. You can recreate this 15-second video using two photos.
The 2010 TV interview gave us a distinct interaction that remains popular today. One man tells another, "Because I know somewhere deep down in my heart... I still love you", and the second man begins to cry heavily. You can take this exact sequence and put two people you know into the two roles. To do this, you provide two photographs to the tool. The tool places the faces and clothes from those photographs onto the bodies in the meme. The background changes to a cosy living room in the evening. The final result is a 15-second vertical 9:16 MP4 video. You can take this MP4 and upload it directly to TikTok, Reels, or Shorts without needing to crop or adjust the dimensions.
Where does the "somewhere deep down in my heart" meme come from?
The original scene comes from a 2010 TV interview. During the broadcast, two men sit across from each other in a serious setting. One man looks directly at the other and delivers the line. He says, "Because I know somewhere deep down in my heart... I still love you". Immediately after he finishes the sentence, the second man bursts into tears. The crying is sudden, intense, and very loud.
This specific sequence of a serious confession followed by an ugly cry is the entire structure of the meme. When you recreate it, the generated video follows this exact pattern from start to finish. There is no music in the background to distract from the interaction. You only hear the generated voice speaking the English words, followed by the loud sobbing of the second person. The scene always features exactly two people. You cannot add a third person to the room, and the site does not animate pets. The focus remains entirely on the two individuals you select.
What do you need to prepare before making the somewhere deep down in my heart video?
Before you start the process, you need to find two photographs on your device. The tool requires exactly two photos to generate the scene. Each photo must show only one person. If you upload a group photo, the tool will not know which face to use for the animation.
The photos must meet specific file requirements to be accepted by the system. You can upload files in JPG, PNG, or WebP formats. The maximum file size for each photograph is 10 MB. If your photos are larger than 10 MB, you must resize or compress them before uploading.
The quality of the faces in the photos determines the quality of the final video. You need to choose photos where the person has a clear, well-lit face. Shadows across the face, blurry images, or hands covering the face will result in a poor video generation. The tool reads the face and the clothes from the photograph. The person in the final video will wear the exact same clothes they wear in the photo you upload. The original background of your photo does not matter at all, because the tool replaces it entirely with a cosy living room in the evening.
How do you choose the right photos for the two roles?
The meme has two distinct roles: the person who cries and the person who says the line. When you make yours, you must assign one photo to the first position and one photo to the second position.
The first photo you upload becomes the person who cries. This is the person who listens to the confession, puts a hand over their mouth, and then bursts into tears. The second photo you upload becomes the person who says the line. This person speaks the English words directly to the first person.
If you put the photos in the wrong order, the roles will be reversed. Swapping the photos swaps the roles completely. You should decide in advance which person you want to cry and which person you want to deliver the confession. Both people will end up sitting in the cosy living room in the evening, but their actions are strictly tied to whether they are photo one or photo two.
What is the step-by-step process to generate the video?
The process of generating the video happens directly on the website. Signing in to the site is free. Choosing your photos and uploading them to the platform is also free. However, there is no free trial video available. You must pay to generate the final MP4.
One video costs $5. This is a one-time purchase. There is no subscription required to use the tool, meaning you only pay when you want to make a video. The payment is handled securely through FloatPass. When you pay the $5, you receive 1,250 platform credits. These 1,250 credits pay for exactly one video generation. The validity of these credits follows the FloatPass terms.
Once you sign in, upload your two photos, and pay the 1,250 credits, the generation process begins. You do not need to keep the page open while it works. The video arrives in a few minutes. When it is ready, it stays in the "My videos" section of your account. You can download the 15-second vertical 9:16 MP4 from there to your phone or computer.
Sometimes, a generation might fail due to a processing error. If a generation fails, the site gives your 1,250 credits back to your balance. You can then try again with the same photos or choose different ones without paying another $5.
How does the 15-second timeline break down?
The final video is exactly 15 seconds long. It is formatted as a vertical 9:16 MP4, which fits the standard dimensions for TikTok, Reels, and Shorts perfectly. The 15 seconds are divided into five specific shots that match the timing of the original 2010 TV interview.
Here is the exact breakdown of the timeline:
| Time | Action |
|---|---|
| 0 to 4.9 seconds | The confession begins. The second person starts speaking. |
| 4.9 to 5.9 seconds | The first person puts a hand over their mouth in anticipation. |
| 5.9 to 8.3 seconds | The second person says, "I still love you". |
| 8.3 to 15 seconds | The first person breaks down into an ugly cry with loud sobbing. |
Throughout this timeline, the generated voice speaks the English line clearly. There is no music playing over the voices or the crying. The audio consists entirely of the spoken words and the loud sobbing, ensuring the focus stays on the dramatic interaction between the two people.
What common mistakes should you avoid with the somewhere deep down in my heart meme?
When people try to make this video, they often make a few specific errors. Knowing these in advance saves time and ensures you get the video you want.
First, do not upload photos of animals. The site does not animate pets. It requires human faces to map the expressions of speaking and crying. If you upload a picture of a dog or a cat, the generation will not work correctly.
Second, do not upload photos with multiple people. The scene has exactly two people, one per photo. If photo one has three people in it, the tool cannot isolate the correct face for the crying role. Ensure there is only one person per photo.
Third, do not expect the tool to change the clothing. Each person keeps the face and clothes from their photo. If you want the person to wear a suit in the video, you must upload a photo of them wearing a suit. The tool only changes the background to the cosy living room in the evening.
Fourth, remember that the audio is fixed. The generated voice will always say "Because I know somewhere deep down in my heart... I still love you" in English. You cannot change the language of the spoken audio, even though the website itself is available in English (en) and Chinese (zh). You also cannot add background music through the site.
Finally, be aware of the payment structure. Signing in and choosing the photos are free, but there is no free trial video. You must purchase the 1,250 platform credits for $5 through FloatPass to see the result. If you encounter any issues, you can reach out via email at contact@deepdowninmyheart.com.
Where else do people look to make this video?
People searching for ways to edit photos into videos often look at several different platforms. You might see people mentioning domains like capcut.com or makeviral.ai when discussing video trends. Other places people also look for meme generation or video editing include getstarrd.app, kling.ai, imgflip.com, and kapwing.com. We do not make claims about the features or prices on those sites, but they are common destinations for people interested in creating short vertical videos for TikTok, Reels, and Shorts.
Frequently asked questions
How much does the somewhere deep down in my heart video cost?
One video costs $5, which equals 1,250 platform credits. This is a one-time purchase with no subscription, and payments are processed through FloatPass.
Can I get a free trial for the somewhere deep down in my heart meme?
No, there is no free trial video. Signing in and choosing the photos are free, but generating the video requires the 1,250 credits.
What happens if I upload the photos in the wrong order?
The first photo is the one who cries, and the second is the one who says the line. Swapping the photos swaps the roles in the video.
What audio plays in the somewhere deep down in my heart video?
A generated voice says "Because I know somewhere deep down in my heart... I still love you" in English, followed by loud sobbing. There is no music in the background.
Can I use a picture of my dog?
No, the site does not animate pets. You must use photos with exactly one person per photo, featuring a clear, well-lit face.