FAQ and Prompt Guide: Luma Dream Machine

🚀 Getting Started with Dream Machine

Welcome to Dream Machine, we put together this page to help you get started with Dream Machine and answer any questions that come up along the way. This is a living document as we continue to grow and learn. Please share any feedback and questions with us in our Discord channel! https://discord.com/invite/lumaai

How to prompt?

Image to Video

• Upload a JPG, JPEG or PNG image by clicking the image icon on the generation page, you can also choose to provide a prompt to describe the scene.
• With “Enhance Prompt” enabled, you just need to describe what you want to happen in the scene. For example: <image of a car> + “*a red car driving on a road*”
  • It’s especially helpful to describe the types of motions that you want to happen in the video
• If you cannot get the desired output, you can try without “Enhance prompt”. This requires describing the image + what you want to happen in the scene.
  • You might see no or low motion with unenhanced prompts. Try to include motion cues in your prompt if necessary.
• In either case, it helps to describe the important details in the image as much as possible, what types of actions are being performed, and how the different parts of the scene should move. E.g., A woman dancing ballet, her arms gracefully extending upward as she twirls around, the dancers behind her bowing low to the ground

Text to video

• A great way to start to get good results is to use “Enhanced Prompt”. By enabling this you can use short prompts like “a teddy bear swimming in the ocean” and the Dream Machine will automatically generate additional descriptions to help create a more detailed generation.
• If you are not able to get the expected results, try without “Enhance Prompt”. Unenhanced prompts need to describe the contents of the scene and the desired action (3-4 sentences). Ex: “*In a somber, nostalgic style, a young man sits on a tree stump in a forest, the warm autumn leaves surrounding him. He wears a brown jacket, dark shirt, and blue jeans, his fingers deftly moving along the fretboard of an acoustic guitar.*”
• In general, you may get better results by being more specific about:
  • Camera motion: *“A dramatic zoom in”, “An FPV drone shot”*
  • Actions and motion: *“A teddy bear swimming with its arms and feet as the turbulent water splashes all around”*
  • Object features: *“A white teddy bear wearing sunglasses with soft fur texture”*
  • Setting and background: “A beautiful cloudy sunset near a Caribbean beach”

---

❓ FAQ - Frequently Asked Questions

1. Why am I not getting the video back/stuck with “in queue”

   The community has been experimenting with Dream Machine. Since launch, we have added extra capacity to scale to meet the demand. To make sure everyone has the best experience possible, we are dynamically limiting and relaxing the number of free generations, and to help those who need it the most - prioritizing paid accounts. We will keep on working to improve everyone's experience and are immensely grateful for your patience and interest.

2. What would a subscription get me?

   • Higher priority generations when there is a queue
   • Commercial use rights
   • Watermark-free downloads (Introduced to all new videos from 6/17 6AM PST)

3. Why are the free credits subtracted when I subscribe? It was a miscommunication on our end and we have fixed that issue now! Paid tiers now get the 30 free generations + 120 paid generations, and this applies to previous paid users as well.

4. I am getting “Error generating video, please try again. (Network Error)?

   If you are doing image to video, this might be because the image type you uploaded is not supported, or your image file is too large. Try converting the image to a jpg, jpeg or png, and reduce the size of the image. We are working on improving this error message and appreciate your feedback.

5. Is it possible to generate more than 5 seconds/add negative prompt/change aspect ratio?

   It's not yet supported. We are working on additional controls for upcoming versions of Dream Machine. We appreciate all of your feedback. If you have anything that you would like to see in upcoming versions please share in our Discord at https://discord.com/invite/lumaai

6. Will it be available on Discord?

   Dream Machine is currently only available on https://lumalabs.ai/dream-machine

7. Is it possible to use the videos from Dream Machine for commercial use?

   The free tier doesn’t grant permission for commercial use, as defined in our Terms of Service https://lumalabs.ai/legal/tos. You can lift this limitation by subscribing to any of the paid tiers.

8. Why am I not getting my generation?

   
   

   If it’s a new generation, you might need to refresh the page or click on the video to see it. If refreshing the page doesn’t work, please send a help ticket in our Discord to let us know.

   *There was a small outage between 10:00-11:00 PM PST on 6/13, causing some of the generations submitted during those times not showing up. We are working on bringing back those videos, and will provide an update here.

9. Does the new “Extend” feature cost credit?

   Yes. The “Extend” button will extend your video for 5 more seconds and will cost 1 generation credit

10. How can I get access to the early features?

    We will have our Dream Machine editor available soon! Sign up here and stay tuned. https://docs.google.com/forms/d/e/1FAIpQLScgsh3Cf2VXf8k7k9iVd-nP5NZk-TSRfq4Tkvh7inv8kS9Ulg/viewform

11. Do unused credits roll over to the next month?

    No, credits will be reset one month from your signup (free tier) or subscription (paid tiers)