Custom AI workstations · Southern Idaho + remote
Custom workstation builds for local AI—spec, assembly, drivers, storage, and a stack you can repeat. Want NAS-style shares, a Flask service, or VMs on the same box? We plan that with the hardware up front so nothing fights for disk, RAM, or cooling later.
Setup is quoted as a project. You can hand off and self-operate. TPS-managed care is optional—only if you want us on contract after delivery. Fleet and desk IT stays on managed IT, not this offering.
Simple breakdown
The quoted project is planning, parts, assembly, OS, and the stack we agree on—including local AI tooling. No monthly plan required. Optional TPS workstation care is described later on this page.
Engineering for the quoted workstation build—hardware, OS, and software—before any optional monthly care. Distinct from the recurring rows in the management table.
Drivers, CUDA or CPU paths, inference runtime layout, quantization choices where appropriate, and dependency pinning.
Storage planning for checkpoints, embeddings, and logs—so growth does not silently fill the disk.
Least-privilege defaults, network exposure choices, and service accounts suited to a local inference workstation.
Written steps for restart, update, and safe experiments—whether you self-operate or add a management tier.
You can hire TechHand for a finished workstation and software setup, pay the quoted project, and operate it yourself. Month-to-month care is optional—only when you want an active TPS agreement on an AI workstation care tier.
One-time (no subscription required)
Consultation, parts list, assembly, OS install, drivers, storage layout, local AI runtimes, and any agreed extras (for example NAS-style shares, Hyper-V or other VM hosting, or a Flask service)—plus smoke tests and written handoff. Priced by quote after we understand goals, models, and what else that box should do.
Optional subscription
If you want TechHand to monitor, patch, and support the same workstation after handoff, that is an active TPS contract on Essential, Business, or Complete AI workstation care. Tiers set response style and which maintenance tasks are in scope—see the comparison table below. Skip this entirely if you prefer self-service operation.
Model licenses, third-party model terms, and legal use of outputs remain your responsibility unless a separate written engagement says otherwise. We do not imply HIPAA, export, or other certifications unless explicitly contracted and documented.
Only if you want TechHand on contract after handoff: ongoing care for one primary AI workstation we built (or explicitly agreed to cover). Monthly pricing is quoted per tier—there is no public rate card here.
Quoted / month
Light-touch care when you mostly run the box yourself—patched, watched, and reachable on business hours.
Highlights
Confirm tier and SLAs in your TPS agreement.
Quoted / month
Adds runtime maintenance, scheduled model assistance, priority response, and alignment with MSP ticketing when you also run under managed IT.
Highlights
Confirm tier and SLAs in your TPS agreement.
Quoted / month
Maximum hands-on care: after-hours escalation, restore drills, quarterly reviews, and priority treatment for model changes.
Highlights
Confirm tier and SLAs in your TPS agreement.
These rows apply only when you choose a TPS-managed care plan for the one workstation on the agreement. Exact SLAs and hours are defined in your TPS statement of work—not on this marketing page.
| Capability | Essential AI Care | Business AI Care | Complete AI Care |
|---|---|---|---|
| One designated AI workstation under plan | |||
| Periodic stack health checks (inference path, disk, thermals) | |||
| OS security patching coordination for AI workstation | |||
| GPU / chipset driver update planning | |||
| Inference runtime & dependency maintenance | — | ||
| Model add / swap assistance with documented rollback | — | ||
| Priority scheduling for model changes | — | — | |
| Backup of configs & critical model / embedding data | |||
| Backup restore drills on agreed cadence | — | — | |
| Remote support — business hours | |||
| Priority remote response window | — | ||
| After-hours escalation for AI workstation | — | — | |
| Incident notes, runbook updates & follow-up tasks | — | ||
| Quarterly performance & capacity review | — | — | |
| Coordination with broader MSP / TNT workflow (when contracted) | — |
Business AI Care includes Essential-level rows plus Business checks above. Complete AI Care includes Business-level coverage plus Complete-only rows. Uptime and outcome guarantees are not implied here—only what we will reasonably perform under contract. For hosted web workloads (not this on-prem workstation), see managed web hosting.
Tell us what you want to run, what else that machine might do (NAS, VMs, Flask, etc.), and whether you only need a quoted build and setup or also an optional TPS care tier. We will answer with a clear scope: parts and labor, software stack, handoff—and management only if you ask for it—plus paired web / workflow work if that is part of the job.
Beyond the inference box
Local models are only useful if your team can reach them safely—through internal tools, queues, approvals, or a public site that calls your stack on your terms. When you need setup, build, manage, or all three, we can scope the same engagement to include custom web and Flask-based workflow software alongside the workstation: one plan path, clear ownership, and the same TPS relationship for what stays on contract.
Examples: a staff portal that submits jobs to a local model, dashboards for review and audit trails, integrations with email or line-of-business systems, or a staged handoff from prototype to production. Details live on our web development & Flask applications page—including policy-aware and regulated-style work when that is part of your requirement.
Typical combined scope
We can quote local AI commissioning together with application work and, where it fits, managed hosting for the web tier—so you are not coordinating three vendors for one workflow. Optional ongoing care for the AI workstation is summarized in the care tiers and comparison table above; app and site care follow whatever support or hosting agreement we write for those layers.