Some users report having their accounts banned within 30 minutes of subscribing while creating legitimate promotional videos. This occurs without warning or clear explanation. Combined with the no-refund policy, users lose both access and money immediately.

Content moderation is arbitrary and lacks transparency. Videos approved one day get flagged the next with nearly identical edits - changing an image or fixing pronunciation triggers rejection. Users left guessing what triggers moderation, creating frustration and wasted time. Healthcare, biotech, and medical users report being blocked entirely without explanation.

Customer support has been criticized for hiding behind policy links and warnings rather than offering real solutions. Users describe it as dealing with an automated system rather than a team that values paying customers. Technical issues and upgrade inquiries get slow responses.

There is no self-service way to close or delete your account. Users must contact support and wait for them to process it. This creates unnecessary friction for users wanting to leave the platform.

Despite using the platform for only two days, Synthesia refuses refunds citing their strict no-refund policy. Users requesting refunds within the stated 15-day consumer refund period report being refused. Support agents refuse to transfer to supervisors. Customers lose money without receiving working service.

Video credits expire monthly with no carryover. For consultants and teams with variable workloads, this creates waste during slow months. If credits could bank even for two months, it would fit better with how consultants actually work. The system isn't designed for fluctuating production schedules.

The platform is expensive for big teams, with costs escalating significantly when adding seats and workspaces. API access requires higher-cost plans. Enterprise contracts involve custom pricing that often surprises organizations after initial quotes.

Every video generated on Synthesia requires human review before download, adding 1-2+ days to the process. This defeats the purpose of using AI for quick and efficient video production. Users expecting instant turnaround are frustrated by mandatory waiting periods.

Avatar quality varies significantly, with facial features often not matching spoken words. Lip-sync accuracy is criticized across multiple platforms. Some avatars fall into the 'uncanny valley' - realistic enough to be unsettling but not enough to be natural. Expressions and gestures feel stiff.

Users developing their own custom avatars face stringent rules that reject images for simple issues like brightness problems. The process is frustrating when legitimate business use cases are blocked by overly strict automated filters.

Voices can still sound robotic on longer narrations or when nuanced emotion is needed. The text-to-speech quality struggles with complex emotional delivery. Shorter clips sound better than extended scripts.

The built-in editor is intentionally simple, meaning limited timeline control, transitions, and audio mixing. Advanced edits require exporting and using external tools, fragmenting the workflow that was supposed to be all-in-one.

Users lack complete control over the gestures, expressions, and movements of virtual presenters. Avatars sometimes feel stiff, especially when expressive delivery is needed. The preset animations limit creative control.

Common technical issues include videos getting stuck during processing (especially at 99% completion) and failing to render entirely. Some users report 80% of videos getting stuck. Resolution issues cause videos to become blurry, requiring redo.

Users experience occasional render queues during peak times, adding unexpected delays to video production. Cloud-based rendering means you're dependent on server availability, which can vary.