Product promos, social media content, explainers, news roundups, brand videos, and more. Wavemaker supports five video aspect ratios: landscape (16:9), portrait (9:16), square (1:1), feed (4:5), and tall portrait (2:3), for YouTube, TikTok, Reels, feed, and Pinterest. The same product also makes downloadable audio (Music, Voice, Spot ads, Podcasts) and platform-ready static image ads.

Short videos usually finish in minutes; longer multi-scene productions take longer. Scenes generate in parallel, and you watch real progress the whole way — every scene, take, and QC verdict, with a live cost estimate before anything spends.

Yes. Just tell it what to change: “make the intro shorter,” “change the background music to something upbeat,” “replace the image in scene 3.” Wavemaker understands context and makes precise edits without regenerating the entire video.

We don’t claim ownership of what you generate, and subject to our terms you may use it for any lawful purpose. As with all AI-generated media, the legal status of AI output can vary, and you are responsible for reviewing and clearing output before you publish it. See our Terms of Service for the full details.

We do not use your prompts or uploads to train our own AI models, and we don’t sell your data. To create your video, your inputs are processed by third-party AI providers in real time, subject to their own terms. See our Privacy Policy for details, including the providers we use.

Yes. Use the Platform API (/runs, /discover) or connect agents via Bearer MCP at $PLATFORM_API_URL/mcp. Worker https://api.wavemaker.io/mcp is retired (HTTP 410). API access is included on paid plans that offer keys.

Credits are consumed when generating videos, static image ads, and downloadable audio. Longer videos and richer AI imagery, voiceover, and music use more; Audio mode scales with purpose, length, and variants. Credits are reserved up front and anything unused is refunded automatically when the run finishes. You can track every charge in your dashboard.

Yes. A workflow is a reusable production pipeline: generation blocks and review gates wired together. Start in Create, describe it in plain English, import a ComfyUI graph (curated packs run natively; unsupported nodes land in a residue report), then publish it to Explore. Shared workflows always load and run: blocks are platform-hosted and versioned.

Not on the live platform. Org BYOK (attach your own xAI/Kling/Gemini/… keys and pay a 20% orchestration fee) was a legacy kernel feature and has no platform-api successor yet — see BYOK. Runs today use Wavemaker-metered GPU + partner calls.

The Training Studio (also: LoRA training) trains custom models from your own material: build a moderated dataset from uploads or your own generations, train an image LoRA (30 credits), video LoRA (90 credits), WAN dual-expert pair (135 credits), or voice clone (30 credits), pick the epoch checkpoint that nails it, and prove the winner with blind A/B evaluations. Failed training jobs refund automatically.

Yes. Open published workflows can be Tried and Remixed. Paid creator premiums stay off in production until Stripe marketplace approval; you can still publish for Try/Remix.

Start creating for free. Describe your video and sign up when you hit generate to claim starter credits, enough for your first short video. Need more? Purchase additional credits from your dashboard.