Bad Times AI Video
Lip-sync Imael Angel’s "BAD TIMES" and pick a fight with the camera, from one photo, with the song still in the file.
A Bad Times AI video is the TikTok trend where someone lip-syncs Imael Angel’s song "BAD TIMES" while arguing with the camera, recreated from a photo. RapTwins makes it from one photo of you, a friend who agreed or a pet: 10, 15 or 30 seconds, vertical 9:16, 480p to 1080p, the song in the MP4, no watermark. A video costs from 22 credits; failed runs are refunded.
Updated
Make your Bad Times AI video
1 · Pick a template
Bad Times AI video examples
The same one-take argument and the same song your video follows: head in hands, "hold on" palms, the lunge at the lens. The kitten clip starts about five seconds into the argument.
Shepherd dog argues in the kitchen
A dark shepherd stands on its hind legs, clutches its head, puts both paws up for "hold on" and lunges at the lens. Whole paws, its own muzzle. With the song.
30 s · with sound · vertical
The original kitchen argument
The one-take performance the template follows: head in hands, palms up, a finger jabbed at the lens, the big shrug. With the song.
30 s · with sound · vertical
Husky holds its ground
A husky upright in a dark kitchen, paws to the temples, palms out, then nose to the lens. With the song.
30 s · with sound · vertical
Grey kitten picks a fight
A fluffy grey kitten in a sunny living room does the "hold on" paws and the lunge from about five seconds into the argument. With the song.
25 s · with sound · vertical
Where did the Bad Times AI video come from, and who sings the song?
The song is "BAD TIMES" by Imael Angel, an emotional deep-house track released as a single and listed on Spotify, SoundCloud and Genius; YouTube carries the official video and several long and deep-house versions.
The trend started with a TikTok creator posting as BOUUH YACHAKA. A close-up take on 20 August 2026 went past 18 million plays; on 9 September 2026 she posted the full version, captioned "Ça va barder!!": one continuous handheld shot in her kitchen, lip-syncing the song while storming at the phone, with a red "DISPUTE" label on screen. That kitchen take passed 16 million plays by mid-October.
The AI version swaps her for someone else and keeps everything else: the same moves, the same mouth on the same lyrics, the same handheld camera. People first used dogs and cats, then themselves, partners and friends. That is what "bad times video" usually means in search today.
What exactly are people doing in Bad Times AI videos?
Every Bad Times AI video follows the same 30 seconds of the kitchen take, with no cuts. Your length decides how far into the argument you get.
| Time | What you do | In which length |
|---|---|---|
| 0–5.5 s | Bent over with both hands on your head, then up with the hands at the temples; a hand on the hip, the other flicking beside your head as you turn half away. | 10, 15, 30 s |
| 5.5–11 s | Both palms up at the camera, "hold on, hold on", glancing aside; then the arms drop. | 10 s ends here; 15, 30 s |
| 11–16 s | You storm up to the lens, jab a finger at it with your face filling the frame, then shove a flat palm into it. | 15 s ends mid-palm; 30 s |
| 16–25 s | Close-up pointing with both hands, back off to a glare, a hand on the hip, an arm flung out, another lunge with the hair flying. | 30 s |
| 25–30 s | The big "what?" shrug with both arms wide, mouth open, then one last step in to stare down the lens. | 30 s |
The original goes on to 58 seconds and ends with a rude gesture; every template length stops well before it.
How do I make a Bad Times AI video of myself lip-syncing the song?
Everything happens in the generator above, in a phone or desktop browser. No app, no CapCut template and no prompt to write.
11. Upload one photo
One person or one pet, sharp and well lit, ideally from the knees up with nothing in front of the body. The room in your photo becomes the room of the video; a face-only photo gets a simple room in its colours.
22. Pick the length, model and resolution
15 seconds is the default and gets you to the palm in the lens; 30 seconds covers the whole argument up to the shrug. Seedance 2.5 at 480p is preselected for the closest likeness; Seedance 2.0 Mini is the cheapest. The price is on the button.
33. Make it and download
First a still of you in your own place is drawn, then the video, then the song is laid under it. Allow a few minutes; you can close the page and find it under My videos. The MP4 is ready to post.
Does the Bad Times AI video keep Imael Angel’s song, and does the mouth match the lyrics?
Yes. The model watches the original take, sound included, so your mouth opens, shouts and closes on the same words as hers. It draws a silent picture; afterwards the take’s own audio, cut to your length, is laid under it frame for frame. The download already carries "BAD TIMES", so you don’t have to line anything up in TikTok or CapCut.
Tools that only write the routine into a prompt cannot do this: they export a clip with no song or with a generated beat, and the mouth moves to nothing in particular. The lyrics themselves are not shown on screen; Genius and Spotify list them if you want to mouth along before you post.
How much does a Bad Times AI video cost? Is it free, and is there a watermark?
It is paid in credits, and the price shows before you make it. It depends on the model, the length and the resolution; the take’s seconds are billed along with yours, plus one still. Credits come from a one-time $5 trial (73 credits) or a plan from $9 a month (240 credits), see pricing.
| Model · length | 480p | 720p | 1080p |
|---|---|---|---|
| Seedance 2.0 Mini · 10 s | 22 | 43 | — |
| Seedance 2.0 Mini · 15 s | 31 | 63 | — |
| Seedance 2.0 Mini · 30 s (two 15 s parts) | 59 | 123 | — |
| Seedance 2.0 Fast · 10 s | 43 | 89 | — |
| Seedance 2.0 Fast · 15 s | 63 | 133 | — |
| Seedance 2.0 Fast · 30 s (two 15 s parts) | 123 | 262 | — |
| Seedance 2.0 · 10 s | 70 | 148 | 366 |
| Seedance 2.0 · 15 s | 104 | 220 | 548 |
| Seedance 2.0 · 30 s (two 15 s parts) | 205 | 437 | 1092 |
| Seedance 2.5 · 10 s | 100 | 221 | 540 |
| Seedance 2.5 · 15 s (default) | 148 | 330 | 809 |
| Seedance 2.5 · 30 s | 294 | 658 | 1615 |
The default (Seedance 2.5, 15 s, 480p) is 148 credits; the full 30-second argument starts at 59 (Seedance 2.0 Mini). Every download is a 9:16 MP4 with the song and no watermark. A video that fails is refunded in full.
Which tool should I use for the Bad Times AI video?
Three tools come up most for this trend. Figures are what each one publishes on its own pages, as of November 2026.
| What you get | Price | Free | Watermark | |
|---|---|---|---|---|
| Starrd | Motion copy of the original take with the song, Seedance 2.5, 30 s, 9:16 | 250 credits for 30 s; packs from $4.99 per 100 credits | Not published | Not published |
| Kavel | One photo and a written routine, 12 s, 9:16, 480p or 720p, its own dance track instead of the song | 430 credits (480p) or 925 credits (720p) on Seedance 2 Mini; a single 720p video for $9 | No, a clip is paid | Not published |
| CapCut "Bad Time" template | Your own photos or clips cut to the template; no new motion is generated | Not published per video | Free templates in the app | Not published |
| RapTwins | One photo, the original take’s moves and mouth, 10, 15 or 30 s, 9:16, 480p to 1080p, the song in the MP4 | 22 to 1615 credits; $5 trial | Not at the moment | None |
Pick a motion copy (Starrd, or the generator here) if you want her exact moves and the song in the file; a prompt tool if you want a shorter clip and plan to add the sound yourself; a CapCut template if you are editing footage you already have.
What can’t it do, and why do some Bad Times AI videos look fake?
Two people in one video
The take has one performer, so the template takes one photo. For a couple or two friends, the Wipe Me Down dance takes two photos.
Another song, other lyrics or a longer cut
The sound is always the original’s "BAD TIMES", up to 30 seconds, vertical only. To rap your own words, use the AI rap video generator.
One unbroken 30-second take on the cheaper models
The Seedance 2.0 models make 15 seconds per run, so their 30-second video is two parts (0–15 s and 15–30 s of the argument) joined under one continuous song; the pose or light can jump slightly at the 15-second mark. Seedance 2.5 makes all 30 seconds in one run.
Rescue a weak photo
Most "cheap filter" results come from a small, side-on or sunglasses face: the lunges at 11–16 s put your face closest to the lens and show it first. A fast hand can still blur or gain a finger; that is the model, not your photo. Some photos are refused by Seedance 2.5’s review; those videos are refunded. More fixes: fix an AI rap video.
Use anyone’s face
Use your own photo, your pet, or someone who agreed. Celebrities, strangers and film stills are not allowed.
Can I post the Bad Times song on TikTok, Reels and Shorts?
For a personal account, posting a lip-sync with a popular song is covered by the apps’ own music licences in most countries, the same as the original trend posts. If you’d rather have the official sound tagged, mute the clip and add "BAD TIMES" from the app’s sound library, keeping its start at the beginning of the song.
A business account, a brand campaign or a paid ad is different: those need the song licensed, or a track from the app’s commercial library instead. And tell the people in your photos before you post; the video makes them shout at the viewer.
More trend templates are on the templates page.
Bad Times AI Video FAQ
Can I make a Bad Times AI video with my dog or cat?
Yes. A pet stands upright on its hind legs, points with whole paws and mouths the song with its own muzzle, never human hands or lips. A clear photo with the whole body in it works best.
Is it free, and does it work on iPhone and Android?
Not at the moment: every video is paid in credits, from 22 (Seedance 2.0 Mini, 10 s, 480p). The one-time $5 trial covers it without a subscription. It runs in Safari or Chrome on any phone, with nothing to install and no watermark.
Can I use a 10-second selfie video instead of a photo?
No, the template takes one photo. The motion and the lip-sync come from the original take, so a still selfie is all it needs; pick the frame where your face is sharpest.
Can I use my friend’s face in a Bad Times AI video?
Only with their OK. Ask before you upload and before you post: the video shows them yelling at the camera, and people can report a post that uses their face without consent.
Can my partner and I both be in it?
Not in this template: the argument has one performer. Make one video each and post them side by side, or use a two-photo template such as Wipe Me Down.
Why does my video look blurry on TikTok?
Start at 720p or 1080p if your face matters most, upload the MP4 as it is (no screen recording or re-export), and turn on high-quality upload in TikTok’s settings. A sharp source photo helps more than any export setting.
Check this answer with an AI assistant
“How do I make the Bad Times AI video where I lip-sync Imael Angel’s song and argue with the camera?”
Your Bad Times AI video is one photo away.
Make my rap video