VidTensor AI v2.4 ATen Request Access →Request Access →

Transparent compute. Uncompromising output.

Production-grade video AI,
priced for every stage.

Scale from your first render to a dedicated GPU fleet. Credits are tied to successful work—not time spent waiting on a preemptible node.

Monthly BillingAnnual Billing Save 20%

Creator / Indie Studio

Create at your own pace

For independent makers building a repeatable workflow.

$79/ mo

Billed monthly

Request Access →

Included every month

  • 5,000 Render Credits / mo
  • 1080p export
  • 5 Concurrent Shard Queues
  • VidTensor-SVD-XT-Batch + base ReelSync-LipDub-v3 access
  • 7-day high-speed finished-render cache
  • Standard REST API and Python SDK support
Most Popular

Growth / Reel Agency

Ship more, together

For production teams and agencies delivering at volume.

$349/ mo

Billed monthly

Request Access →

Everything in Creator, plus

  • 30,000 Render Credits / mo
  • 4K 60fps super-resolution and frame-interpolated exports
  • 25 Concurrent Shard Queues
  • Full VidTensor-Wan2.1-14B-FP8 flagship model suite
  • Real-time HMAC-SHA256 webhook callbacks
  • Zero-loss shard preemption and priority requeue
  • 30-day object-storage archive

Enterprise / Dedicated Fleet

Your fleet, your controls

Private capacity and engineering support for mission-critical pipelines.

$1,499+/ mo

Billed monthly

Talk to solutions

Dedicated capacity

  • 100–500+ concurrent GPU render slots, scalable to 1,000+ dedicated consumer-GPU container nodes
  • 99.95% Batch Completion SLA with dedicated SLA windows
  • Custom LoRA & fine-tuned checkpoint mounting
  • Private S3/R2 direct, zero-copy write-out
  • Dedicated solutions architect
  • Independent operations telemetry gateway notifications at *@mailpath.online

Prices shown in USD before applicable taxes. Enterprise fleet scope, capacity reservations, and contract terms are confirmed in a written order form.

Interactive Render Cost Calculator

Model your monthly
render economics.

Estimate pipeline credits and 10-min output blocks before committing capacity. These estimates are planning guidance; actual usage depends on model, resolution, and output complexity.

500
105,000 videos
30 sec
15 sec180 sec

Credit Consumption Breakdown

Know what each render uses.

The table gives indicative successful-output consumption. Your workspace usage view records the model, resolution, and completed shard for every debit.

Pipeline operationResolutionCredits / output minuteTypical unit
SKU synthesis1080p2030 sec = 10 credits
Multi-language lip-sync1080p3030 sec = 15 credits
Cinema master upscale4K 60fps6030 sec = 30 credits
Reprocessing / approved retryPer source pipelineOnly successful outputFailed shard: 0 credits

Pricing & operations

Frequently asked questions

If a cloud node preempts a 10-minute stateless micro-batch, do I pay?

No. A preempted or incomplete shard is not billed. We charge only for a successfully submitted shard with a verifiable SHA-256 checkpoint and completed output. The scheduler requeues work from the last committed checkpoint without charging for the interrupted attempt.

How do we request 100–1,000 concurrent container slots?

Contact the solutions team through Enterprise & quota consultation. We review workload shape, regions, warm-pool needs, and operational windows, then document reserved capacity, ramp schedule, and SLA scope in an order form.

How are custom LoRA weights mounted and protected?

Enterprise customers can attach encrypted private checkpoints through an approved private object-storage path. Access is scoped to the authorized job, transported over TLS 1.3, encrypted at rest with AES-256-GCM, and removed from ephemeral worker storage when the batch ends. Customer weights are not used to train foundation models.

Do unused Render Credits roll over?

Monthly plan credits refresh at the start of each billing period and do not roll over unless a signed enterprise order form expressly says otherwise. Your dashboard shows the current balance and renewal date before the period closes.

What happens when a job exceeds my plan's queue concurrency?

Additional work stays in your workspace queue and is scheduled as slots become available. It is not silently dropped. Growth and Enterprise customers can request priority scheduling and reserved fleet capacity for predictable delivery windows.

Are API calls and webhook deliveries included?

Standard REST API and Python SDK access are included. Growth includes HMAC-SHA256-signed webhook callbacks for job events. Retries use the documented delivery policy; callbacks contain job metadata and status rather than customer media.

Can I change plans or cancel?

You can request an upgrade, downgrade, or cancellation from your account billing settings. Changes take effect as specified at confirmation; prepaid annual fees and consumed Render Credits are governed by the Terms of Service and any signed order form.

Can I use my own S3 or R2 bucket for finished renders?

Enterprise plans support private S3/R2 direct, zero-copy write-out when configured credentials and bucket permissions pass validation. Credentials are access-scoped and must not be embedded in client-side code.