• Choosing a model

  There's no single best model - each has strengths: Prompt adherence - how literally the model follows instructions Identity consistency - how well it preserves faces/characters from references Lip-sync and language support - varies a lot between video models Speed vs. quality - "fast" model variants trade fidelity for turnaround We regularly add new models and update existing ones. When something matters (a client deliverable, a character's identity), test 2–3 models wiFew readers

• How long do generations take?

  Typical times vary widely by model, resolution, duration, and current provider load. Images usually take seconds to a couple of minutes; videos commonly take several minutes, and long or high-resolution videos can take considerably longer. This is normal across the industry - the same model on other platforms takes similar time. If a generation seems stuck for a very long time, refresh the page and check Assets - completed results are saved even if the progress view didn't update. If it genuineFew readers

• Generating images in Studio

  Open Studio and select an image model. Write your prompt. Use Wise Enhance if you'd like the AI to expand and improve it. Optionally attach reference images (see the References category). Choose your aspect ratio and quality, then click Generate. Results appear in your Studio feed and are saved to Assets automatically. Hover an image to download, favorite, or reuse it.Few readers

• Generating videos

  Video generation works like image generation, with a few extra options depending on the model: duration, resolution, start/end frames, and reference inputs (images, video, or audio on models that support omni references). Tip: models differ significantly in strengths - lip-sync quality, motion, prompt adherence, and supported languages vary by model. If a result disappoints, try the same prompt on a different model before rewriting it.Few readers

• Generating audio

  Studio supports text-to-speech, sound effects, and music generation. Select the audio mode, choose a voice or style, and describe what you need. Generated audio is saved to Assets and can be used as reference input in video models that accept audio references.Few readers

• Language and lip-sync support

  Spoken-dialogue quality depends on the model's training data. Most video models are strongest in English, Mandarin, Japanese, and Korean. Less common languages — and regional variants (e.g. European vs. Brazilian Portuguese) - may produce inaccurate pronunciation or unusable lip-sync on some models, regardless of how the prompt is written. If dialogue in your language comes out wrong, this is usually a model limitation rather than a prompt problem. Try a different video model - language covFew readers

• Quality settings and credit cost

  Most image models offer resolution tiers (e.g. 1K / 2K / 4K) and most video models offer resolution options (e.g. 720p / 1080p). Higher settings cost more credits and take longer. The credit cost of your current selection is always shown on the Generate button before you run it.Few readers