Frequently Asked Questions

Frequently Asked Questions

Find answers to common questions about Recaster licensing, technical requirements, troubleshooting, and content guidelines.

Browse the categories below or use the search function to find answers to your questions. Click any question to expand the answer.

Licensing

Free includes all local processing features at no cost: face swapping (InSwapper), face enhancement (GFPGAN), video upscaling (Real-ESRGAN, SwinIR 2x/4x), training, face editing, and more. Studio ($99 one-time) adds cloud GPU processing via Vast.ai, live preview streaming, multi-session support, advanced models (Ghost, BlendSwap, CodeFormer, GPEN, RestoreFormer++, Real-ESRGAN 8x), and priority support.

No. Studio is a one-time purchase of $99. You own it forever with no recurring fees. Cloud GPU costs (Vast.ai) are separate and billed directly by the provider based on your usage.

Yes. Your license can be activated on up to 3 devices simultaneously. You can deactivate a device through the app settings or the Customer Portal to free up a slot for a new device.

Your license key was sent to your email when you signed up. Check your inbox (and spam folder) for an email from Recaster. You can also log into the Customer Portal at recaster.studio/portal/license to view your key or request a resend.

Yes. You can upgrade at any time by purchasing a Studio license from the website. Your existing projects, settings, and trained models are fully preserved. The Studio license key replaces your Free key when activated.

Technical

For local processing, a dedicated NVIDIA GPU with 4 GB+ VRAM is recommended. Face swapping works with as little as 2 GB VRAM, while training benefits from 6-12 GB. If you do not have a powerful GPU, Studio tier users can process on cloud GPUs via Vast.ai. CPU-only processing is supported but significantly slower.

Yes. Recaster runs natively on macOS 12+ including Apple Silicon (M1/M2/M3/M4). Face swapping and enhancement work well on Mac. Real-ESRGAN upscaling is CoreML-accelerated. Note that DFL training is very slow on Mac without NVIDIA CUDA, so remote training is recommended for Mac users.

Currently only NVIDIA GPUs are supported for GPU acceleration. AMD and Intel GPU users will fall back to CPU processing automatically. Studio tier cloud GPUs provide an alternative for users without NVIDIA hardware.

The application itself requires about 500 MB. AI models are downloaded on first use and total approximately 500 MB. Projects vary in size: a typical video project may need 1-10 GB depending on video length and quality. We recommend at least 50 GB of free space for comfortable working.

Troubleshooting

If you are using CPU processing (no NVIDIA GPU detected), performance will be significantly slower. Check Settings to verify your GPU is being used. Close other GPU-intensive applications. Studio users can offload heavy processing to cloud GPUs for the best performance.

Try clearing the settings folder and restarting. On macOS: ~/Library/Application Support/Recaster/. On Windows: %APPDATA%\Recaster\. On Linux: ~/.config/Recaster/. After clearing, you will need to reactivate your license and accept the BIPA consent dialog.

Ensure you are copying the entire key including the dot separator. The format is data.signature with exactly one dot. Check for extra whitespace or line breaks. If the problem persists, log into the Customer Portal to view your key or request a resend. You can also ask for help on Discord.

Make sure you have the latest NVIDIA drivers installed (version 535+). Restart your computer after updating drivers. Verify your GPU with nvidia-smi in the terminal. Ensure your GPU has at least 2 GB VRAM. AMD and Intel GPUs are not supported for acceleration.

Content Guidelines

Recaster automatically blocks processing when a detected face appears to be under 25 years old. This is a safety measure required for legal compliance. The policy uses the FairFace classifier which works with age ranges rather than exact ages, so a margin above 18 is necessary to prevent minors from being processed.

Recaster may not be used to create: non-consensual intimate imagery (NCII), child sexual abuse material (CSAM), content intended for election interference or political misinformation, or any content designed to deceive, harass, or harm individuals. Violations may result in license revocation.

Yes. You must obtain appropriate consent or have legal rights to use any person's likeness before processing. This applies to all individuals including actors, public figures, and private individuals. Recaster requires BIPA consent acknowledgment on first launch.

Legitimate professional use cases include: film and TV production (stunt doubles, de-aging, digital doubles), creative content creation with consent, educational projects, visual effects research, and artistic expression where appropriate consent and disclosure are maintained. All AI-generated content should be labeled as such.

Still Have Questions?

Can't find what you are looking for? Our community is active and ready to help. Join the Discord server to ask questions, report issues, or share your work.

Ask on Discord