Build workload containers
Connect
SPIRITT
RunPodRunPod
Build the control layer for GPU work
Configure RunPod GPU templates, endpoints and approved clusters with controlled deployment workflows. Hand SPIRITT the outcome and let your agent carry it through.
What you can do with RunPod.
Choose hardware deliberately
Compare returned GPU specifications, availability and prices against the workload and approved budget.
Keep deployment inputs controlled
Configure templates, private registry access and secrets without exposing credentials in code or reports.
Build a GPU operations console
Build a private workload setup app with approved resource choices, known-pod desired state and configuration evidence.
Works with your other apps.
Get started in three steps.
01
Open your workspace
Start a SPIRITT workspace and describe the operation you want to run around GPU options, pod configurations, templates, endpoints and registry credentials.
02
Connect RunPod
Connect RunPod, choose the account and records in scope, and authorize the companion apps your mission needs.
03
Hand over a mission
Your agent builds the system, works from your GPU options, pod configurations, templates, endpoints and registry credentials, and follows unresolved actions to their outcome.
Hand over a mission.
Choose a mission. Make it yours.
Put RunPod to work.
Build the control layer for GPU work
Questions
RunPod, with SPIRITT.
01What can SPIRITT do with RunPod?+
Configure RunPod GPU templates, endpoints and approved clusters with controlled deployment workflows. Compare returned GPU specifications, availability and prices against the workload and approved budget.
02Can SPIRITT build an app around RunPod?+
Build a private workload setup app with approved resource choices, known-pod desired state and configuration evidence. Your agent builds the interface, connects the source records, and keeps the workflow running.
03Does this connection run inference or stop pods?+
It configures templates and serverless endpoints, creates clusters and reads a known pod. Inference-job submission, pod stopping and job-result retrieval are not exposed here; those need the workload’s separately verified execution path.
04How does SPIRITT handle cost and secret boundaries?+
Resource creation follows an approved budget and configuration. Serverless templates use a zero-gigabyte volume setting. Secrets and registry credentials remain in approved secure stores, and financial account fields require elevated access.
05Can this RunPod workflow run on a schedule?+
Yes. SPIRITT can revisit your GPU options, pod configurations, templates, endpoints and registry credentials on an agreed schedule, keep a durable record of changes, and follow outstanding work through connected apps. It checks returned results before treating an action as complete.
06How do I get started with RunPod?+
Open a SPIRITT workspace, connect RunPod, and hand over a mission. Choose the records, desired outcome, and actions your agent can take.