VidMuse is the ABSOLUTE WORST
Update: I will say this: VIDMUSE IS A SCAM! On 09/13/2026, I cancelled my subscription and wrote a review detailing the horrible job they did with an unusable music video. Today (09/25/2026), I just discovered that they still billed me! The sad part about this is that I can't reach support because they refuse to respond. Legal action needs to be taken by everyone on TrustPilot to shut this piece of crap platform down! VidMuse is the ABSOLUTE WORST! I'll even go as far as saying that it's a scam! I initially had 12100 credits for a 4 minute 25 second music video. I uploaded the music, uploaded the character, had specific instructions for the script and emphasized accurate lip syncing. What I got was a video that looks like an oil painting with a few photorealistic scenes and a lip sync that matches maybe a total of 25 out of the 265 seconds. To fix the lip sync issue, I was charged basically all of my credits, and the problem was still not resolved. I contacted support twice about the lip sync issue. They gave me a total of 4500 free credits to fix the issue. In addition, I purchased an additional 4800 credits. Guess what happened? NOTHING!! The AI chatbot explained the lip sync issue as this: "Seedance 2.5 cannot produce frame-accurate phoneme sync from an audio reference. Regenerating these 22 clips with the same model and adding anamorphic/grain styling will improve the look, but will not guarantee accurate lip sync. The only way to get true word-level sync is OmniHuman V1.5 (audio-driven avatar model) — which renders as a portrait/talking-head format rather than the cinematic multi-shot format you designed." I opted to go the OmniHuman route and THAT failed also, while eating away at my credits. In the meantime, I made multiple attempts to fix the cartoon look. The chatbot said that they were fixed, but my timeline said otherwise. I was then told this about the lip sync: "If H3 is still producing mismatched lip sync on your spoken-word track, this is hitting the current ceiling of AI video generation technology. No amount of re-rolling or parameter adjustment will reliably fix it — the models simply don't yet have perfect spoken-word phoneme accuracy for all voices and cadences." So basically, VidMuse can advertise perfect lip sync, but they actually are limited. According to the chatbot, The real issues: Lip sync — The H3 model has a systematic offset problem where it aligns to the start of the audio clip rather than the track position. This affects nearly every clip. Support has not responded in 3 days despite multiple escalations. Photorealism — Despite multiple rounds of prompt rewrites, the AI generation models continue producing stylized/illustrated results on many clips. This is a genuine capability limitation — the models don't always honor photorealism instructions consistently. 17,000 credits spent — A significant portion of that was on regenerations caused by model failures, not your creative decisions. Just in case VidMuse is reading this, here's what was documented for support: Urgent escalation submitted. Here's exactly what was documented for support: 3 days with no response — flagged as unacceptable given the scale of the issue Lip sync root cause clearly stated — H3 model aligns to audio clip start, not track offset; support's own recommended workaround was tested and confirmed non-functional Photorealism failure documented — 10 specific clips listed that still render as cartoonish despite 5+ rounds of prompt rewrites Full credit spend breakdown — ~17,000 credits across 6 rounds of generation, majority caused by model failures not user changes Four specific demands made — 24-hour personal response, technical disclosure on lip sync capability, full credit audit and refund, honest disclosure of platform limitations I'm stuck with a video I can't use, I've used 17000 credits, and I'm still being asked to top up more credits to see if what they couldn't previously fix can be fixed. This is a total ripoff, and I wish there were some way to report this money grab.