Introduction
Luma Dream Machine 2.0, powered by the Ray 3 model, is one of the leading AI video generation platforms in 2026. Whether you are evaluating the platform for the first time or have just signed up, you likely have questions that the marketing pages do not fully answer.
This FAQ compiles the most common questions about Dream Machine 2.0, organized by topic. Answers are based on the platform’s specifications as of March 2026 and reflect both official documentation and practical user experience.
Video Specifications
Q: What is the maximum video length Dream Machine 2.0 can generate?
A: The maximum clip length depends on your plan:
- Free plan: Up to 5 seconds
- Standard plan: Up to 10 seconds
- Pro plan: Up to 10.5 seconds (256 frames at 24 fps)
- Enterprise plan: Custom (typically matches Pro ceiling, with longer durations available on request)
The slight difference between Standard (10 seconds) and Pro (10.5 seconds) reflects the Pro plan’s access to the full 256-frame context window.
Q: What resolutions are supported?
A: Dream Machine 2.0 supports the following output resolutions:
| Resolution | Aspect Ratio | Available On |
|---|---|---|
| 720p (1280×720) | 16:9 | All plans |
| 1080p (1920×1080) | 16:9 | Standard, Pro, Enterprise |
| 1080p (1080×1920) | 9:16 (vertical) | Standard, Pro, Enterprise |
| 1080p (1080×1080) | 1:1 (square) | Standard, Pro, Enterprise |
| 1080p (2560×1080) | 2.39:1 (CinemaScope) | Pro, Enterprise |
| 4K (3840×2160) | 16:9 | Enterprise (limited availability) |
The Free plan is restricted to 720p. CinemaScope 2.39:1 aspect ratio is exclusive to Pro and Enterprise plans.
Q: What frame rate does Dream Machine 2.0 generate?
A: The native generation frame rate is 24 fps, the standard cinematic frame rate. Pro and Enterprise plans offer frame interpolation to 48 fps for smoother motion, with a slight reduction in fine detail sharpness as a trade-off.
Q: What file formats can I export?
A: Available export formats:
- MP4 (H.264): Default, universally compatible
- MP4 (H.265/HEVC): Smaller file size at equivalent quality
- WebM (VP9): For web embedding
- ProRes 422: Available on Pro and Enterprise plans, for professional editing workflows
- GIF: Available for clips under 5 seconds
Q: Can I generate videos longer than 10 seconds?
A: Not in a single generation pass. For longer content, you have two options:
- Storyboard mode (Pro plan): Plan and generate a sequence of connected clips with automatic scene transitions and character consistency. Each individual clip is still limited to 10.5 seconds.
- Manual stitching: Generate multiple clips and edit them together in external software (DaVinci Resolve, Premiere Pro, Final Cut, etc.).
Generation Modes
Q: What is the difference between text-to-video, image-to-video, and video-to-video?
A:
- Text-to-video: Describe a scene in natural language. The model generates the video entirely from your text description.
- Image-to-video: Upload a reference image (photo, illustration, 3D render). The model animates the image with motion described in your text prompt.
- Video-to-video: Upload an existing video clip. The model restyled or extends it while preserving the original structure and motion.
Credit cost comparison (5-second, 1080p, Enhanced quality):
- Text-to-video: ~15 credits
- Image-to-video: ~12 credits
- Video-to-video: ~18 credits
Image-to-video is cheapest because the model has more structural guidance from the reference frame.
Q: What is “character lock” and how does it work?
A: Character lock (Pro and Enterprise plans only) allows you to preserve a character’s visual appearance across multiple generated clips. After generating an initial clip, you can “lock” a character’s face, clothing, body proportions, and overall appearance. Subsequent generations that reference the locked character will maintain visual consistency.
The system uses CLIP-based embedding anchoring combined with a lightweight LoRA adapter trained on-the-fly from the first generation. It works best when:
- The character is clearly visible in the initial generation
- The locked character occupies at least 15 % of the frame
- Style and lighting remain broadly consistent between clips
Character lock is not perfect — subtle variations in facial detail can occur, especially with significant changes in lighting or viewing angle between clips.
Camera Controls and Prompt Tips
Q: What camera movements does Dream Machine 2.0 support?
A: The platform supports both preset-based and prompt-based camera control:
Presets (accessible via the camera panel):
- Dolly in / dolly out
- Crane up / crane down
- Orbit left / orbit right
- Handheld (simulated natural shake)
- Steadicam (smooth tracking)
- Static (locked camera)
Prompt-based (interpreted from text):
- “Slow push-in on the subject’s face”
- “Low-angle tracking shot following the car”
- “Bird’s eye view rotating clockwise”
- “Dutch angle tilt as the character falls”
Q: What camera parameters can I control?
A:
| Parameter | Range | Plan Availability |
|---|---|---|
| Focal length | 14 mm – 200 mm (simulated) | Standard, Pro, Enterprise |
| Aperture (depth of field) | f/1.2 – f/16 | Standard, Pro, Enterprise |
| Shutter angle (motion blur) | 45° – 360° | Pro, Enterprise |
| Camera speed | Slow / Medium / Fast | All plans |
Q: What are the best prompt writing tips for Dream Machine 2.0?
A: Based on community experience and Luma’s own guidance:
1. Start with the camera, then the scene:
- Good: “Slow dolly-in, a woman in a red dress walks through a sunlit market”
- Less effective: “A woman in a red dress walks through a sunlit market, camera slowly moves in”
2. Specify lighting explicitly:
- Good: “Golden hour side lighting from the left, warm color temperature”
- Less effective: “Nice lighting” (too vague)
3. Include material and texture cues:
- Good: “Matte black ceramic vase on a white marble countertop”
- Less effective: “Black vase on a counter”
4. Describe motion with direction and speed:
- Good: “Gentle wind from the right causes her hair to sway slowly”
- Less effective: “Windy hair”
5. Use cinematic terminology: The model responds well to professional terms: “shallow depth of field,” “anamorphic bokeh,” “rack focus,” “motivated lighting,” “practical light sources,” “lens flare.”
6. Avoid contradictions:
- Problematic: “A bright sunny day with dark dramatic shadows and overcast sky”
- Better: Pick one lighting scenario and commit to it.
7. Keep prompts under 200 words: Extremely long prompts tend to lose coherence. The sweet spot is 40–100 words with specific, non-contradictory instructions.
Commercial License
Q: Can I use Dream Machine 2.0 content commercially?
A: Commercial use rights depend on your plan:
| Plan | Commercial Use | Conditions |
|---|---|---|
| Free | No | Personal and non-commercial only |
| Standard | Yes | Standard commercial license |
| Pro | Yes | Extended commercial license |
| Enterprise | Yes | Full commercial license with indemnification |
Q: What does the commercial license cover?
A: The Standard and Pro commercial licenses grant you the right to use generated content in:
- Social media posts and advertisements
- YouTube videos (including monetized)
- Client work and freelance deliverables
- Product marketing materials
- Presentations and pitch decks
The license does not cover:
- Reselling generated content as stock footage without modification
- Using generated content in a way that implies endorsement by real individuals
- Generating content that violates Luma’s Terms of Service (deepfakes, non-consensual imagery, etc.)
Q: Does Enterprise indemnification protect me from IP claims?
A: Enterprise plans include a legal indemnification clause where Luma agrees to defend and indemnify the customer against third-party intellectual property claims arising from the use of generated content, subject to the terms of the Enterprise agreement. Specific coverage varies by contract — consult Luma’s sales team for details.
Q: Do generated videos have watermarks?
A: Free plan videos include a small Luma watermark in the lower-right corner. Standard, Pro, and Enterprise plans produce unwatermarked output.
API Access
Q: Does Dream Machine 2.0 have an API?
A: Yes. API access is available on Pro and Enterprise plans.
- Pro plan: REST API + Python SDK. Rate-limited to approximately 50 concurrent requests.
- Enterprise plan: REST API + Python SDK. Higher rate limits, dedicated infrastructure options, and SLA guarantees.
The API supports all generation modes (text-to-video, image-to-video, video-to-video) and returns results via webhook or polling.
Q: What is the API response time?
A: API generation times mirror the web platform:
- Standard quality, 5-second clip: ~30–60 seconds
- Enhanced quality, 10-second clip: ~90–180 seconds
- Cinema quality, 10-second clip: ~180–300 seconds
Pro plan API requests are prioritized over Standard plan web requests.
Content Moderation
Q: What content is not allowed?
A: Luma’s content policy prohibits:
- Non-consensual imagery of real individuals
- Explicit sexual content
- Content depicting real minors
- Graphic violence intended to glorify harm
- Content promoting illegal activities
- Deepfakes intended to deceive
Q: Are there content restrictions that affect legitimate creative work?
A: Some users report that Luma’s content filters can be overly cautious with:
- Stylized violence common in action/thriller genres
- Artistic nudity in fine art or educational contexts
- Dark or horror-themed aesthetics
- Military or combat scenarios
If your creative work is flagged incorrectly, you can request a content policy review through Luma’s support team. Enterprise plans offer customized content policy configurations for approved use cases (e.g., medical education, security training).
Technical and Account Questions
Q: What browser and hardware do I need?
A: Dream Machine 2.0 is a browser-based platform. Minimum requirements:
- Modern browser: Chrome 110+, Firefox 115+, Safari 17+, Edge 110+
- Stable internet connection (generation is server-side, but previewing requires bandwidth)
- No GPU required on your device — all rendering happens on Luma’s servers
Q: Can I use Dream Machine 2.0 on mobile?
A: The web interface is responsive and functional on mobile browsers, but the camera control panel and storyboard mode are optimized for desktop. For the best experience, use a desktop or tablet.
Q: How do I cancel my subscription?
A: Subscriptions can be canceled from your account settings page. Cancellation takes effect at the end of the current billing period — you retain access to your plan features until then. Unused credits are not refunded and do not roll over after cancellation.
Troubleshooting
Q: My generated video looks blurry or low-quality. What happened?
A: Common causes:
- You are on the Free plan (720p maximum)
- You are using Standard quality mode instead of Enhanced or Cinema
- Your prompt is too vague, causing the model to generate a generic scene
- The subject is too small in the frame, reducing effective detail
Q: The video does not match my prompt. What should I do?
A: Try these adjustments:
- Simplify the prompt — remove conflicting or ambiguous instructions
- Add more specific visual descriptions (materials, lighting direction, colors)
- Use image-to-video mode with a reference frame for more control
- Generate 3–4 variations and select the best one
- Break complex scenes into simpler individual clips
Q: Generation is taking much longer than expected.
A: Generation times increase during peak hours (typically 9 AM–6 PM US Pacific time). Pro plan users receive priority processing. If generation exceeds 10 minutes for a 5-second clip, contact support — the job may be stuck.
References
- Luma Labs: https://lumalabs.ai
- Luma Dream Machine: https://lumalabs.ai/dream-machine
- Luma pricing: https://lumalabs.ai/pricing
- Luma API documentation: https://docs.lumalabs.ai
- Luma Terms of Service: https://lumalabs.ai/terms
- Luma content policy: https://lumalabs.ai/content-policy