Name who you are coming for, add one photo, and you get a 12-second vertical video of that person spitting cocky, original bars over a dark AI beat — lip-synced and ready to drop into a group chat or a story. Lyrics are free to draft, the full video costs $3.99 as a one-time purchase with no subscription, and every verified account gets a free 6-second watermarked preview first. Most diss track and diss rap generators hand you a wall of lyrics and leave the performance to you; this diss rap generator hands you the performance. It is built for the things friends actually argue about: whose fantasy team collapsed, who never replies, who reheats fish in the office microwave. One rule before you start: the target has to be someone who will laugh when they see it.
How to make a diss track
- Keep the Diss Track template. It is already selected in the generator above. It sets the scene — a dim recording booth, one hard spotlight, mic in hand — plus the dark beat and the cocky delivery.
- Add one or two photos. Upload clear photos of adults who agreed to appear. One photo makes that person the rapper. Two photos put the first one on the mic and the second reacting behind them, which is the version people screenshot.
- Write the topic. Optional, but this is where the diss comes from. Give the target's name and two or three habits, hobbies or running jokes — "Dan, still plays the same one deck, calls it strategy, lost three in a row". Leave it empty and the AI writes a generic call-out instead.
- Generate. The AI writes the bars, raps them over the beat and renders the lip-synced video. Want every word to be yours? Open "Write my own lyrics", draft or type the verse, and the video raps exactly that.
Which photos work best
- One person per photo. A group shot gives the model two faces to choose from, and it will pick the wrong one.
- Front-facing, eyes open. A head turned far to the side or tipped back loses the face halfway through the clip.
- Even light, no sunglasses. Hard shadow across half a face, hats low over the eyes and heavy filters all cost you detail.
- Frame from the chest up. A full-body photo taken from across a room leaves too few pixels in the face to rebuild.
- Adults who agreed to appear. No photos of minors, and no faces you do not have permission to use.
- JPEG, PNG or WebP, under 10 MB each. Straight off the camera roll is fine.
What makes a diss land
- Specific beats savage. "You still owe me for the Uber in March" lands harder than any generic insult. Three real details will carry the whole verse.
- Punch at choices, not at people. Habits, takes, hobbies, the team they support, the way they load a dishwasher. Not bodies, not families, not where anyone is from.
- One punchline and one hook. Twelve seconds is about 30 words. Pick your best joke, repeat the hook, and let the rest go.
- End on a mic drop. The last line is the one that gets quoted, so make it the hardest one.
- Ask for a clean version if this is going to a family chat or a work channel — just say so in the topic.
Keep it a joke, not harassment
The line is simple: you are making something the target can laugh at, not something that hurts them. So, in plain words — no slurs, no threats of violence, no doxxing, no sexual content. Only use photos of people who agreed to appear, and only of adults. The lyrics model is told to keep the bars off appearance, body, family, race, religion, disability and sexuality, and we remove videos that cross that line. If someone tells us a video of them was made without their consent, we take it down. The details are in DMCA and takedowns and the Terms of Service.
How it compares to filming it yourself
Three ways to end up with a diss track video, and what each one asks of you:
- Template apps you edit yourself. You film the footage, cut it to the beat and source a track you are allowed to use. Full control over the look, and an evening of work before anything is postable.
- General-purpose AI video tools. You write the prompts, generate the clip, and then fix the timing and the lip sync yourself, usually over several attempts. Flexible, but the mouth matching the words is your problem.
- This generator. One photo and a template. The lyrics and the beat are generated, the lip sync is done for you, and you download a 12-second vertical MP4. $3.99 for one 480p video, one-time, after a free 6-second preview.
Diss track or roast battle?
They look similar and they are not the same video.
- Diss Track — one person on the mic, calling out a target who is not answering back. Use it when the joke is aimed at somebody. This is the template on this page.
- Roast Battle — two people face to face in an underground club, trading jabs. Use it when both of them are in on it and both get a turn.
- Hype Duo — the friendly version: two people on a concert stage hyping each other up instead of swinging. Good for a birthday, a win or a send-off.
Both of the others are one click away on the AI rap generator page. Want the battle bars written first? The AI rap battle generator gives you a free 8-bar verse to start from, and the text to rap generator raps any text you paste in.
What it costs
Lyric drafts are free within guest limits and need no account. Credits go on the video only: each verified-email account gets one play-only, 6-second 480p watermarked preview. A downloadable 12-second 480p video is 30 credits and 720p is 60. Credit packs are one-time, with no subscription: $3.99 for 30 credits, which is exactly one 480p video, then $9.99 for 100, $29.99 for 350 and $79.99 for 1,000. Credits do not expire. The lyrics used are saved with every video, so you can copy them for a caption.
No app to install, no charge for a failed render
- Runs in your browser. Nothing to install on a phone or a laptop; close the tab and we email you when the video is ready.
- No charge if a render fails. Credits — or your free preview eligibility — come back automatically.
- Paid videos download clean. The 12-second MP4 carries no watermark. Only the free 6-second preview is watermarked, and that one is play-only.
- Your files leave when you say so. Delete the project and its photos, audio and videos are removed from our storage right away.
Can I diss over a real beat or sound like a famous rapper?
No. Every track uses an original AI beat chosen by the template. You cannot upload a song or pick a different beat, and the voice is not an imitation of any specific artist's recording or sound. That is deliberate: a diss track with somebody else's instrumental under it is their copyright, not yours, and a voice clone of a named rapper is their likeness. Rights holders can request removal through our DMCA process.
Ready to open fire? Write your target into the generator above, or start from the AI rap video generator home page.
Frequently asked questions
Can AI write a diss track about my friend and make his face rap it?
Yes. Name the target, add two or three real details about them, upload a clear photo of them, and you get a 12-second vertical video of that face rapping original diss bars over a dark AI beat, lip-synced. Lyric drafts are free, so you can read the verse before you render anything.
Do I need one photo or two?
One is enough. One photo makes that person the rapper on the mic. Two photos put the first person on the mic and the second one reacting behind them, which is the version people screenshot.
Can it sound like a famous rapper or use a real beat?
No. Every track uses an original AI beat chosen by the template, you cannot upload a song or pick a different beat, and the voice is not an imitation of any specific artist’s recording or sound. A diss track with somebody else’s instrumental under it is their copyright, and a voice clone of a named rapper is their likeness.
Is it allowed to roast someone?
Only if it stays a joke the target can laugh at. Use photos of adults who agreed to appear, punch at choices rather than at people, and keep out slurs, threats, doxxing and sexual content. The lyrics model is told to avoid appearance, body, family, race, religion, disability and sexuality, and we take down videos that cross that line or that were made of someone without their consent.
How long does it take and what does it cost?
A 12-second render usually takes about one to two minutes, plus queue time when the site is busy, and we email you when it is ready. One 480p video is 30 credits and 720p is 60. The $3.99 pack is exactly 30 credits, which is one 480p video; larger packs are $9.99 for 100 credits, $29.99 for 350 and $79.99 for 1,000, all one-time with no subscription. Credits do not expire, and a failed render returns them automatically.
Can I download it for TikTok?
Yes. Paid videos download as a vertical 9:16 MP4 with no watermark, ready for TikTok, Instagram Reels, YouTube Shorts or a group chat. The free 6-second preview is watermarked and plays on the site only.
What's the difference between a diss track and a roast battle?
A diss track is one person on the mic calling out a target who is not answering back, so use it when the joke is aimed at somebody. A roast battle is two people face to face in an underground club trading jabs, so use it when both of them are in on it and both get a turn. Hype Duo is the friendly version: two people on a concert stage hyping each other up. If you only want the jokes as text, the free [AI roast generator](/ai-roast-generator) writes them without a video.
What happens to my photos and videos?
Uploaded photos and the generated audio and video stay in your project in our own cloud storage, readable only through your account. Delete the project and its photos, audio and videos are removed from our storage right away.
Last updated: 2026-11-01