Luma Dream Machine Credits Explained: Web, iOS, API and Production Spend

Short answer: Luma Dream Machine credits only make sense after you choose the route. Web and iOS plans use monthly app credits, top-ups extend an active paid app plan, relaxed mode changes speed after fast credits, and API work uses a separate developer budget. Treat those as different wallets before estimating production spend.

Start with the credit boundary

Luma credits are not one universal wallet. Dream Machine web and iOS subscription credits pay for app-based image and video creation, top-up credits extend a paid app account, and API balances sit in a separate developer route. Buyers should choose the surface first, then estimate credits inside that surface.

That separation matters because Luma publishes more than one official pricing surface. Dream Machine support pages explain web and iOS plans, monthly credits, top-ups, fast mode, relaxed mode, and per-generation credit costs. Luma's broader pricing page also presents the current Luma workspace, team routes, shared credits, and model cost tables. The controlling source is the route shown at checkout for the account that will do the work.

The practical rule is simple: do not divide one monthly price by a generic number of videos. A credit estimate needs the route, model, duration, resolution, quality mode, edit type, retries, upscales, and whether the work is human-led in the app or automated through an API.

Web and iOS plan routes

The web route is the cleaner starting point for most solo creators who will create inside Dream Machine from a browser. Luma's Dream Machine support page lists a web Free plan with limited monthly credits, non-commercial use, lower priority processing, draft resolution, and watermarks. The paid web route then moves through Lite, Plus, Unlimited, and Enterprise.

Lite is the low-cost web test route, but it remains a constrained creative plan. It includes a 3,200 monthly credit allowance, full Ray3 access, higher priority processing, and 4K with up-res, while Luma still describes it as non-commercial and watermarked. That makes it a trial and learning route, not the safest production publishing route.

Plus is the first obvious web route for commercial output because Luma lists 10,000 monthly credits, commercial use, no watermarks, high priority processing, full Ray3 access, and 4K with up-res and HDR. Unlimited keeps 10,000 monthly fast credits but adds relaxed mode for ongoing generation after fast credits run out. Enterprise adds a larger monthly credit bundle, highest priority processing, commercial use, no watermarks, no training input/output data, and sales-led pricing.

The iOS route follows a similar creative workflow but uses its own plan prices and allowances. Luma's support page lists iOS Free with 250 monthly credits, iOS Lite with 3,200 credits, iOS Plus with 10,000 credits, and iOS Unlimited with 10,000 fast credits plus relaxed mode. Choose iOS when mobile app billing and creation are part of the workflow; compare against web before assuming the same plan name has the same total cost.

Generation costs that move the budget

Credit burn changes sharply with model, duration, and output settings. Luma's Dream Machine support page says Photon image generation costs 4 credits per image and runs in batches of four, while Image V2 costs 4 credits at 720p and 8 credits at 1080p. Image Modify V2 is much heavier, and video costs rise further with model, resolution, HDR, EXR, and video-to-video work.

For video, use the per-second or per-clip tables as a planning tool, not as a promise that every accepted asset will cost the same. Luma's pricing page lists Ray3.14 text-to-video and image-to-video at 20 credits per second for 720p and 80 credits per second for 1080p. The same table shows HDR and video-to-video paths costing materially more, so a 10-second approved clip can hide several expensive discarded attempts.

Top-ups and reset rules are part of the cost model. Monthly subscription credits reset with the billing cycle and do not roll over. Top-up credits are different: Luma's credit-system page says they start at $4 for 1,200 credits, are valid for 12 months, roll over, and are used after monthly credits. They are also tied to an active paid plan, so a buyer should not treat top-ups as a standalone wallet.

Fast mode and relaxed mode affect more than speed. Lite and Plus do not include relaxed mode, so running out of credits means buying top-ups or waiting for the next billing cycle. Unlimited and Enterprise include relaxed mode after fast credits are exhausted, but relaxed generations are lower priority and slower. That can be fine for exploration and background ideation, but risky for deadline-driven production.

API and production boundaries

The API route is a separate purchase decision. Luma's Dream Machine API documentation describes direct image and video generation through API requests, including text-to-video, image-to-video, text-to-image, character reference, image reference, style reference, and image-to-image workflows. Luma's credit pages explicitly say Dream Machine subscription credits do not transfer to the API.

For Dream Machine API budgeting, verify the billing dashboard rather than assuming app credit math. Luma's API changelog describes a credits endpoint that returns the API credit balance in cents USD, which is a useful clue that developer billing and app subscriptions are not the same accounting system. Luma's API terms also say unused API call quotas, request allowances, or credits allocated within a billing period do not roll over and expire at the end of the applicable period.

There is also a separate Luma Agents API pricing route for Uni image generation. Luma Agents API pricing offers pay-as-you-go for prototypes and lower-volume workloads, and Provisioned Throughput for production workloads that need guaranteed capacity and latency. It meters internally in credits at an approximate dollar conversion shown in the dashboard, but publishes per-image prices and monthly provisioned commitments instead of asking buyers to reuse Dream Machine app credits.

That makes the API the right route when generation must run inside software: a product feature, campaign system, creative operations pipeline, internal tool, or backend process with logs and spend controls. It is usually the wrong route when a human can create, review, and export assets manually from the web or iOS app.

Buyer checks before production spend

Start by writing down the exact route: web Dream Machine, iOS Dream Machine, broader Luma workspace, Dream Machine API, Luma Agents API, Team, or Enterprise. Then record which wallet is paying for the work: monthly subscription credits, top-up credits, relaxed mode, shared team credits, API balance, provisioned throughput, or a negotiated enterprise allowance.

Next, estimate the actual creative loop. Count accepted outputs, discarded attempts, alternative prompts, reference-image tests, video-to-video modifies, upscales, reframes, HDR or EXR exports, audio work, and any batch variations. The cheapest prompt example is not a production budget if the real workflow needs multiple tries and higher-resolution finishing.

For team or client work, verify commercial use, watermarks, no-training terms, shared credit controls, usage analytics, SSO, admin ownership, cancellation behavior, and who can buy top-ups. For API work, verify rate limits, latency expectations, failed-generation charges, credit balance reporting, rollover language, data-use terms, and whether provisioned capacity is required before launch.

The safe purchase path is to test on the smallest route that matches the real surface. Use Free or Lite only for learning and non-commercial tests. Use Plus when commercial browser output and 10,000 monthly credits are enough. Use Unlimited when slower relaxed generations are acceptable after fast credits run out. Use Team, Enterprise, or API routes only when governance, shared credits, no-training requirements, automation, capacity, or production integration are more important than the simplicity of the app subscription.