Midnight Atelier

Bernini AI Video Generator for more deliberate short scenes.

Create text-to-video or image-to-video work in one focused studio: pick a Seedance model, set the render boundary, and inspect credits and safety before the task enters the queue.

Direction study

Reference in. A more intentional frame out.

Reference anchoredMotion directedSafety ready
Reference frame
Rendered direction
Reference frameRendered direction
Creative direction

A measured night drive after rain; preserve the vehicle silhouette, deepen the cobalt reflections, and let warm storefront light pass across the frame.

Illustrative direction study using original Bernini Video Lab artwork. The live product currently supports text-to-video and image-to-video.

One creative surface

Start with direction. Keep every decision in view.

The live workspace connects text-to-video, image-to-video, and render review in one simple sequence — not three separate tools.

Original Midnight Atelier desk with a cinema camera and a lit workbench
  1. 01 · Text to video

    Write the motion brief.

    Turn a written direction into a real video task, then choose only the quality, duration, and ratio the selected model supports.

  2. 02 · Image to video

    Anchor the subject or mood.

    Bring one reference image into the brief when a product, material, composition, or mood needs a steadier visual point of reference.

  3. 03 · Render review

    Inspect the handoff before it queues.

    Prompt safety, render settings, expected credits, and task status stay visible instead of disappearing behind one generate button.

Motion brushes, video references, and audio timelines are not presented as live controls until they are connected to the production workflow.

From thought to task

A four-part handoff that keeps creative intent intact.

The flow is intentionally compact: establish the reference, write the instruction, choose the render boundary, then review the result in your workspace.

  1. 01

    Reference

    Start with a frame when the subject, object, material, or environment matters.

    Identity · material · atmosphere

  2. 02

    Prompt

    Describe the camera, pace, scene, and change in language your team can review.

  3. 03

    Render

    Set quality, ratio, and duration within the model options that are actually available.

  4. 04

    Review

    Track the task, inspect its output, and carry the working brief into the next pass.

Original studio direction board

A visible system

A studio flow, not a black box.

Original visual direction sets the tone, while the product stays candid about what it can do, what it costs, and what is screened before a job is created.

  • Reference material can guide image-to-video jobs without claiming absolute identity lock.
  • Every live render is limited to the model settings shown in the generator — up to 1080p and roughly 15 seconds where supported.
  • Blocked requests do not reach a provider or spend credits.

Render credits

Credits that map to real renders.

Choose a one-time pack for focused tests, a monthly pool for steady output, or annual credits for the lowest effective rate. Every live checkout stays connected to the production payment flow.

Prompt moderation before generation
Credits shown before each job
Billing support at support@bernini.app

Bernini Starter Pack

For a small prompt test batch before scaling production.

$15
  • 60 AI video credits, valid for 90 days
  • Credit cost shown before every generation
  • Text-to-video and image-to-video workflows
Popular

Bernini Creator Pack

One-time credits for testing Bernini prompts, reference media, and first paid video workflows.

$39
  • 180 AI video credits, valid for 90 days
  • Task history, R2 storage, and video downloads
  • Tracked provider model updates
  • Private project workspace

Bernini Studio Pack

A larger one-time pack for production experiments and client drafts.

$99
  • 600 AI video credits, valid for 180 days
  • Priority review queue
  • Priority support
  • API access when the public API opens

Clear before you render

Every plan includes credits for online AI video jobs. The estimated cost is visible before generation, and prompts must pass content safety checks before a task is submitted.

Prompts are screened before generation. Blocked or flagged requests do not enter the model and do not consume credits.

If the provider fails, cancels, or returns no usable result, the credit return path is triggered. For payment refunds or billing questions, email support.

Refund and billing emailsupport@bernini.app

Useful boundaries

Questions a working studio asks first.

Straight answers about what is live now, how credits work, and how a task moves through the service.

Bring the next brief into the studio.

Start with the real generator, see the current model boundary, and keep the first render decision in one place.