AI environment subscriptions
Start with a lower-cost managed AI queue, or put your approved workflows on a physical computer dedicated to one business. The difference is visible, priced, and documented before anything is connected.
Conversations are always free. We identify the workload, data path, response needs and human approval boundary before recommending shared or dedicated compute.
The plain-language difference
The shared plan keeps business workspaces and permissions separate, but processing capacity is scheduled among customers. Dedicated plans reserve a physical computer for one organization.
Swipe sideways to compare all three plans.
| What changes | Shared AI Queue | Dedicated AI Computer | Dedicated AI Operations |
|---|---|---|---|
| Physical hardware | Shared across subscribed customers | Reserved for one organization | Higher-capacity node reserved for one organization |
| Work queue | Scheduled shared queue; jobs may wait | Private queue; starts when your machine is available | Private priority queue for multiple approved workflows |
| Workspace | Separate repository, rules and access | Separate repository plus physically dedicated compute | Dedicated compute plus multi-user workflow controls |
| Best fit | Drafting, publishing, routine reports and non-urgent automation | Frequent processing, local-model options and predictable access | Operational systems, multiple users and heavier media or data work |
| Maintenance | Ten Thirty Two manages the shared environment | Scheduled updates and monitoring on the reserved machine | Priority support, scheduled updates and planned shared-queue failover |
| Starting monthly price | $249 | $995 | $1,995 |
Introductory subscription pricing
Every plan includes workflow definition, a documented approval path, access controls, status records and managed updates. Third-party usage is separate and approved before activation.
Shared AI Queue
Private workspace · shared processing hardware
Dedicated AI Computer · Most private hardware
One physical computer · one organization
Dedicated AI Operations
Higher capacity · multiple approved users
Activation: $299 Shared Queue, $1,995 Dedicated Computer or $3,500 Dedicated Operations. Dedicated subscriptions have a 12-month initial term because equipment is acquired and configured for the client.
What is a standard run? One approved workflow pass that stays inside the agreed processing, storage and model budget. Large media, long-running research and bulk jobs are estimated before they enter the queue.
Hardware ownership: Subscription hardware remains Ten Thirty Two property unless a separate purchase agreement says otherwise. Client-owned computers can be evaluated for managed service starting at $495 per month.
Usage outside the plan: Extra shared workflow runs start at $3 each. Storage, SMS, external model APIs, software licenses and unusually large data transfer are billed separately only after approval.
What can live in the environment
The environment is not sold as a magic employee. Each workflow gets a clear input, job, approval boundary and output record.
Transcription, clip preparation, captions, metadata, approval queues and publishing support.
Content drafts, structured updates, search-page preparation, link checks and controlled deployments.
Extraction, classification, summaries, recurring reports and governed internal knowledge assistance.
Approved Mentra Live pilots, observation intake, requests and follow-through through a documented tunnel and processor.
Security without magic words
We document where data starts, where it travels, which model processes it, who can approve an action and what gets recorded. A secure tunnel reduces exposure; it is not a promise that no incident can ever occur.
The business keeps account ownership, multifactor authentication and final customer-facing approval.
Workflows receive only the permissions they need. Passwords do not belong in footage or working repositories.
Local and external model use is identified before launch. External providers remain subject to their own terms.
Updates, backups, recovery expectations and service interruptions are written into the approved scope.
Common questions
Shared Queue customers have separate workspaces but take turns on shared processing hardware. A Dedicated Computer is a physical machine reserved for one organization and does not enter the shared hardware queue.
No. Agreed local models can stay on the machine. A workflow using an external AI model or software provider sends the required data to that provider under its terms. We document the architecture first.
It can be placed at an agreed client location or a managed location, depending on power, connectivity, physical access and recovery requirements.
Subscription hardware remains Ten Thirty Two property unless a separate purchase agreement says otherwise. We can also evaluate client-owned hardware.
Yes. We can move approved repositories, workflow rules and connection records into a dedicated environment after hardware and migration scope are approved.
No. Model APIs, storage, SMS, software licenses and unusually large data-transfer costs are itemized separately or connected to client-owned accounts.
Choose from the work backward
We will map the job and recommend the least expensive environment that can meet the real requirement. The conversation is free.