Skip to content
verify mcp Beta VerifyMCP is currently in beta. If you notice any issues, get in touch and we’ll put it right.

com.mcparmory/runpod

PYPI · MCPARMORY-RUNPOD · 2 COMPONENTS · SCANNED SEP 20

Launch, scale, and manage GPU pods and serverless endpoints across regions

0 this week 60 Trust /100
Trust breakdown (7 categories)

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. How we score → Why this is hard to score →

Supply Chain Security37
  • Malware scan not yet available for this package.Unverified
  • No known CVEs affecting this package version or its production dependencies.Pass
  • Install-script risk not yet assessed.Unverified
  • 2 of 41 dependencies flagged as unhealthy. View diagnostics → Partial
Provenance & Transparency32
Schema Quality & AI Usability69
  • AI-judged instruction clarity (excellent).Pass
  • Context-footprint check failed: tool/resource definitions use about 6888 tokens (~196/item across 35 items; 35 tools + 0 resources), over budget; trim descriptions and params. See how to fix → Fail
  • Usage-examples check failed: none of the tools include examples. See how to fix → Fail
Stability & Change Management83
  • Stability observed for 25 of 30 days with no destabilising changes; credit accrues until the full window elapses.Partial
Tool Coverage100
  • 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
  • 100% of tool parameters carry a description.Pass
Tool Safety100
  • No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
  • All 5 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation.Pass
  • An AI judge read all 35 captured unit(s) of tool text and found none that tries to manipulate the model reading it.Pass
Capabilities100
  • Implements a current MCP spec version (2026-07-28).Pass
Install

How do I install the com.mcparmory/runpod MCP server?

com.mcparmory/runpod runs locally as a PyPI package, launched with uvx mcparmory-runpod. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

pypi · mcparmory-runpod

# add to Claude Code
claude mcp add com-mcparmory-runpod -- uvx mcparmory-runpod
// .cursor/mcp.json
{
  "mcpServers": {
    "com-mcparmory-runpod": {
      "command": "uvx",
      "args": [
        "mcparmory-runpod"
      ]
    }
  }
}
// .vscode/mcp.json
{
  "servers": {
    "com-mcparmory-runpod": {
      "command": "uvx",
      "args": [
        "mcparmory-runpod"
      ]
    }
  }
}
# add to Codex CLI
codex mcp add com-mcparmory-runpod -- uvx mcparmory-runpod
// opencode.json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "com-mcparmory-runpod": {
      "type": "local",
      "command": [
        "uvx",
        "mcparmory-runpod"
      ],
      "enabled": true
    }
  }
}
# add to OpenClaw
openclaw mcp add com-mcparmory-runpod --command uvx --arg mcparmory-runpod
# ~/.hermes/config.yaml
mcp_servers:
  com-mcparmory-runpod:
    command: "uvx"
    args: ["mcparmory-runpod"]
// ~/.netclaw/config/netclaw.json
{
  "McpServers": {
    "com-mcparmory-runpod": {
      "Transport": "stdio",
      "Command": "uvx",
      "Arguments": [
        "mcparmory-runpod"
      ]
    }
  }
}
# add to Vellum
assistant mcp add com-mcparmory-runpod -t stdio -c uvx -a mcparmory-runpod
// mcp.json
{
  "mcpServers": {
    "com-mcparmory-runpod": {
      "command": "uvx",
      "args": [
        "mcparmory-runpod"
      ]
    }
  }
}
Changelog

Every change we have recorded for this component, newest first. Security-relevant changes are always shown. ▲ marks a change for the better, ▼ a change for the worse; unmarked changes are neutral.

  • 20 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 80 to 83. That category is still filling its 30-day observation window: 24 days of observed history at the previous scan, 25 at this one. The score rises as the window fills, whether or not the server changes.

  • 19 Sept 26 −3
    • Stability: pass → 0.80 functional
  • 18 Sept 26 0
    • Stability: 0.97 → pass security
  • 17 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 93 to 97. That category is still filling its 30-day observation window: 28 days of observed history at the previous scan, 29 at this one. The score rises as the window fills, whether or not the server changes.

  • 15 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 87 to 90. That category is still filling its 30-day observation window: 26 days of observed history at the previous scan, 27 at this one. The score rises as the window fills, whether or not the server changes.

  • 13 Sept 26 −14
    • Malware scan: pass → unverified security
  • 12 Sept 26 +12
    • Malware scan: unverified → pass security
    • Stability: pass → 0.80 functional
  • 11 Sept 26 0
    • Stability: 0.97 → pass security
Diagnostics

Diagnostic detail from the automated scan of this channel: what the scanner observed at each step, so you can see exactly where a check passed or failed. It is informational only and never changes the trust score.

Captured 20 Sept 2026 · Analysed pypi/mcparmory-runpod@1.0.1

Provenance No attestation

The registry publishes no build provenance for this version, so there is nothing to verify.

Result No attestation
Ecosystem pypi

Background: How many MCP packages publish verified provenance →

Dependencies 41 packages
Packages resolved 41
Stale 1
No linked repository 1
Tree resolution Complete

Background: SBOMs and build attestations, explained →

MCP tools · 35 exposed · ~6,888 tokens

The tools this component advertises to a client, with an estimated token cost for each. Expand a tool to see its parameters and schema. The per-tool counts are indicative and are not scored directly; the schema's total context footprint is one signal in Schema Quality & AI Usability. A tool's description is untrusted text the model reads on every call, which is what makes this list a security surface and not just an inventory: how tool poisoning works →

Tool Tokens
create_container_registry_auth ~125

Create a new container registry authentication credential with a unique name. This stores the username and password needed to authenticate with a container registry.

NameTypeReqDescription
namestringyesA unique identifier for this container registry credential. Choose a descriptive name that helps you identify which registry or account this credential is for.
passwordstringyesThe password or authentication token for accessing the container registry. This is stored securely and used when authenticating registry operations.
usernamestringyesThe username or account identifier for accessing the container registry. This is paired with the password to authenticate registry operations.

No output schema declared.

No examples provided.

create_network_volume ~133

Create a new network volume in a specified Runpod data center. The volume will be allocated with the specified storage capacity and can be referenced by its user-defined name.

NameTypeReqDescription
dataCenterIdstringyesThe Runpod data center where the network volume will be created (e.g., EU-RO-1).
namestringyesA user-defined name for the network volume. Names do not need to be unique and can be any descriptive label.
sizeintegeryesThe storage capacity to allocate for the network volume, specified in gigabytes. Must be between 0 and 4000 GB.

No output schema declared.

No examples provided.

create_pod ~859

Create and optionally deploy a new Pod with configurable compute resources, networking, and deployment settings. Supports both GPU and CPU Pods across Secure Cloud and Community Cloud environments.

NameTypeReqDescription
cloudTypestringCloud environment for the Pod. SECURE provides dedicated infrastructure with guaranteed availability; COMMUNITY offers lower-cost shared resources. Defaults to SECURE.
computeTypestringCompute type for the Pod. GPU Pods include graphics processors and ignore CPU-related settings; CPU Pods are GPU-less and ignore GPU-related settings. Defaults to GPU.
countryCodesarrayList of ISO 3166-1 alpha-2 country codes where the Pod can be located. If omitted, the Pod can be placed in any country.
cpuFlavorIdsarrayOrdered list of CPU flavor IDs for CPU Pods. The order determines rental priority; earlier entries are attempted first. Ignored for GPU Pods.
dataCenterIdsarrayOrdered list of data center IDs where the Pod can be located. The order determines rental priority; earlier entries are attempted first. Defaults to a global list of 26 data centers.
dockerEntrypointarrayDocker ENTRYPOINT override for the container. Pass an empty array to use the image's default ENTRYPOINT. Defaults to empty array.
dockerStartCmdarrayDocker CMD override for the container startup. Pass an empty array to use the image's default CMD. Defaults to empty array.
envobjectEnvironment variables to inject into the Pod container as key-value pairs. Defaults to empty object.
globalNetworkingbooleanEnable global networking for the Pod. Currently available only for On-Demand GPU Pods on select Secure Cloud data centers. Defaults to false.
gpuCountintegerNumber of GPUs to attach to the Pod. Only applies to GPU Pods. Must be at least 1. Defaults to 1.
gpuTypeIdsarrayOrdered list of GPU type IDs for GPU Pods. The order determines rental priority; earlier entries are attempted first. Ignored for CPU Pods.
imageNamestringContainer image tag to run on the Pod (e.g., runpod/pytorch:2.1.0-py3.10-cuda11.8.0-devel-ubuntu22.04).
interruptiblebooleanCreate an interruptible (spot) Pod instead of reserved. Interruptible Pods cost less but can be stopped anytime to free resources. Defaults to false (reserved).
lockedbooleanLock the Pod to prevent stopping or resetting. Defaults to false (unlocked).
minDiskBandwidthMBpsnumberMinimum disk bandwidth in megabytes per second (MBps) required for the Pod.
minDownloadMbpsnumberMinimum download speed in megabits per second (Mbps) required for the Pod.
minRAMPerGPUintegerMinimum RAM in gigabytes (GB) per GPU for GPU Pods. Defaults to 8 GB per GPU.
minUploadMbpsnumberMinimum upload speed in megabits per second (Mbps) required for the Pod.
minVCPUPerGPUintegerMinimum virtual CPUs per GPU for GPU Pods. Defaults to 2 vCPUs per GPU.
namestringUser-defined name for the Pod. Does not need to be unique. Maximum 191 characters. Defaults to 'my pod'.
portsarrayList of exposed ports in format [port_number]/[protocol], where protocol is either 'http' or 'tcp' (e.g., ['8888/http', '22/tcp']).
supportPublicIpbooleanFor Community Cloud Pods, set to true to request a public IP address. On Secure Cloud, Pods always have public IPs. Defaults to null (may not have public IP on Community Cloud).
templateIdstringUnique identifier of a Pod template to use for creation. If provided, the Pod is created from this template.
vcpuCountintegerNumber of virtual CPUs to allocate to the Pod. Only applies to CPU Pods. Defaults to 2 vCPUs.

No output schema declared.

No examples provided.

create_serverless_endpoint ~458

Create a new Serverless endpoint for running containerized workloads. Configure compute resources (GPU or CPU), scaling behavior, worker limits, and data center preferences to match your inference workload requirements.

NameTypeReqDescription
computeTypestringCompute resource type for workers: GPU for GPU-accelerated inference, or CPU for CPU-only workloads. GPU-related properties are ignored for CPU endpoints and vice versa. Defaults to GPU.
cpuFlavorIdsarrayList of RunPod CPU flavor IDs available for CPU endpoints. Order determines rental preference priority when scaling workers.
dataCenterIdsarrayList of RunPod data center IDs where workers can be deployed. Order determines preference priority. Defaults to all available global data centers.
gpuCountintegerNumber of GPUs per worker for GPU endpoints. Must be at least 1. Defaults to 1.
gpuTypeIdsarrayList of RunPod GPU type IDs available for GPU endpoints. Order determines rental preference priority when scaling workers.
namestringUser-defined name for the endpoint. Does not need to be unique. Maximum 191 characters.
networkVolumeIdsarrayList of network volume IDs to attach to the endpoint. Enables multi-region endpoints to access shared storage across data centers.
scalerTypestringScaling strategy: QUEUE_DELAY scales workers when requests exceed maximum latency tolerance, REQUEST_COUNT scales based on request queue depth. Defaults to QUEUE_DELAY.
scalerValueintegerScaling threshold: for QUEUE_DELAY, maximum seconds a request waits before scaling up; for REQUEST_COUNT, requests per worker ratio. Must be at least 1. Defaults to 4.
templateIdstringyesUnique identifier of the template defining the container image and runtime configuration for workers.
vcpuCountintegerNumber of vCPUs allocated per worker for CPU endpoints. Defaults to 2.
workersMaxintegerMaximum number of concurrent workers. Must be 0 or greater. Set to 0 for unlimited scaling.
workersMinintegerMinimum number of always-running workers. Must be 0 or greater. These workers run continuously at reduced cost even with no active requests.

No output schema declared.

No examples provided.

create_template ~411

Create a new Docker-based template for RunPod Pods or Serverless workers. Templates define the compute environment, Docker image, runtime configuration, and exposed ports for containerized workloads.

NameTypeReqDescription
categorystringThe compute hardware category for this template. Choose NVIDIA for GPU acceleration, AMD for AMD GPUs, or CPU for CPU-only workloads. Defaults to NVIDIA.
dockerEntrypointarrayOverride the Docker image's ENTRYPOINT instruction. Provide as an array of command segments (e.g., ['/bin/bash', '-c']). Leave empty to use the image's default ENTRYPOINT.
dockerStartCmdarrayOverride the Docker image's CMD instruction. Provide as an array of command segments. Leave empty to use the image's default CMD.
envobjectEnvironment variables to inject into the container at runtime. Provide as key-value pairs (e.g., {'ENV_VAR': 'value'}).
imageNamestringyesThe Docker image to use for this template, specified as a registry path (e.g., 'nvidia/cuda:12.0' or 'myregistry.com/myimage:latest'). Required.
isPublicbooleanMake this template visible to other RunPod users. Only applies to Pod templates. Defaults to private (false).
isServerlessbooleanSpecify whether this template is for a Serverless worker (true) or a standard Pod (false). Defaults to Pod (false).
namestringyesA human-readable name for this template. Used for identification in the RunPod UI and API. Required.
portsarrayList of network ports to expose on the Pod. Each port is specified as 'port_number/protocol' where protocol is either 'http' or 'tcp' (e.g., ['8888/http', '22/tcp']).
readmestringDocumentation for this template in Markdown format. Displayed to users who select this template. Defaults to empty.

No output schema declared.

No examples provided.

delete_container_registry_auth ~53

Permanently delete a container registry authentication configuration. This removes the stored credentials and access settings for the specified registry.

NameTypeReqDescription
containerRegistryAuthIdstringyesThe unique identifier of the container registry authentication configuration to delete.

No output schema declared.

No examples provided.

delete_endpoint ~47

Permanently delete an endpoint by its ID. This action cannot be undone and will remove the endpoint from the system.

NameTypeReqDescription
endpointIdstringyesThe unique identifier of the endpoint to delete.

No output schema declared.

No examples provided.

delete_network_volume ~51

Permanently delete a network volume by its ID. This operation removes the network volume and its associated resources from the system.

NameTypeReqDescription
networkVolumeIdstringyesThe unique identifier of the network volume to delete.

No output schema declared.

No examples provided.

delete_pod ~44

Permanently delete a Pod by its ID. This operation removes the Pod and cannot be undone.

NameTypeReqDescription
podIdstringyesThe unique identifier of the Pod to delete.

No output schema declared.

No examples provided.

delete_template ~39

Permanently delete a template by its ID. This action cannot be undone.

NameTypeReqDescription
templateIdstringyesThe unique identifier of the template to delete.

No output schema declared.

No examples provided.

get_container_registry_auth ~53

Retrieve a specific container registry authentication configuration by its unique identifier. Returns the complete details of the requested registry auth.

NameTypeReqDescription
containerRegistryAuthIdstringyesThe unique identifier of the container registry authentication configuration to retrieve.

No output schema declared.

No examples provided.

get_endpoint ~104

Retrieve a single endpoint by its ID. Optionally include details about the template used to create it and the workers currently running on it.

NameTypeReqDescription
endpointIdstringyesThe unique identifier of the endpoint to retrieve.
includeTemplatebooleanWhen enabled, includes detailed information about the template that was used to create this endpoint. Defaults to false.
includeWorkersbooleanWhen enabled, includes information about all workers currently running on this endpoint. Defaults to false.

No output schema declared.

No examples provided.

get_network_volume ~46

Retrieve a specific network volume by its unique identifier. Returns detailed information about the requested network volume.

NameTypeReqDescription
networkVolumeIdstringyesThe unique identifier of the network volume to retrieve.

No output schema declared.

No examples provided.

get_pod ~189

Retrieve a single Pod by its ID with optional related resource information. Use include parameters to expand the response with machine, network volume, savings plans, template, or worker details.

NameTypeReqDescription
includeMachinebooleanWhen enabled, includes details about the machine the Pod is running on. Defaults to false.
includeNetworkVolumebooleanWhen enabled, includes information about any network volume attached to the Pod. Defaults to false.
includeSavingsPlansbooleanWhen enabled, includes details about savings plans applied to the Pod. Defaults to false.
includeTemplatebooleanWhen enabled, includes information about the template used by the Pod, if one exists. Defaults to false.
includeWorkersbooleanWhen enabled, includes Pods that are Serverless workers in the results. Defaults to false.
podIdstringyesThe unique identifier of the Pod to retrieve.

No output schema declared.

No examples provided.

get_template ~130

Retrieve a single template by its ID. Optionally include templates from Serverless endpoints, community-made public templates, or official Runpod templates in the response.

NameTypeReqDescription
includeEndpointBoundTemplatesbooleanWhether to include templates that are bound to Serverless endpoints in the response. Defaults to false.
includePublicTemplatesbooleanWhether to include community-made public templates in the response. Defaults to false.
includeRunpodTemplatesbooleanWhether to include official Runpod templates in the response. Defaults to false.
templateIdstringyesThe unique identifier of the template to retrieve.

No output schema declared.

No examples provided.

list_container_registry_auths ~32

Retrieves a list of all container registry authentication configurations. Use this to view available registry credentials and their settings.

Input schema present but exposes no named parameters.

No output schema declared.

No examples provided.

list_endpoint_billing_history ~327

Retrieve aggregated billing records for your Serverless endpoints, with flexible filtering by location, GPU type, endpoint, and time range. Results can be grouped by endpoint, pod, or GPU type.

NameTypeReqDescription
bucketSizestringTime bucket size for aggregating billing records. Choose from hourly, daily, weekly, monthly, or yearly aggregation. Defaults to daily.
dataCenterIdarrayFilter results to endpoints in specific Runpod data centers. Provide an array of data center IDs (e.g., EU-RO-1, US-TX-3). Defaults to all available data centers.
endTimestringEnd date for the billing period in ISO 8601 format (e.g., 2023-01-31T23:59:59Z).
endpointIdstringFilter results to a single endpoint by its ID.
gpuTypeIdarrayFilter results to endpoints with specific GPU types attached. Provide an array of GPU type names (e.g., NVIDIA GeForce RTX 4090).
groupingstringGroup billing records by endpoint ID, pod ID, or GPU type. Defaults to grouping by endpoint ID.
imageNamestringFilter results to endpoints created with a specific container image.
startTimestringStart date for the billing period in ISO 8601 format (e.g., 2023-01-01T00:00:00Z).
templateIdstringFilter results to endpoints created from a specific template by its ID.

No output schema declared.

No examples provided.

list_endpoints ~81

Retrieves a list of all available endpoints. Optionally include details about the templates used to create endpoints and the workers currently running on them.

NameTypeReqDescription
includeTemplatebooleanWhen enabled, includes template information for each endpoint. Defaults to false.
includeWorkersbooleanWhen enabled, includes details about workers currently running on each endpoint. Defaults to false.

No output schema declared.

No examples provided.

list_network_volume_billing ~188

Retrieve aggregated billing records for your network volumes over a specified time period. Results can be grouped by hour, day, week, month, or year.

NameTypeReqDescription
bucketSizestringThe time granularity for aggregating billing data. Defaults to daily buckets if not specified. Valid options are hour, day, week, month, or year.
endTimestringThe end of the billing period to retrieve, specified as an ISO 8601 datetime string (e.g., 2023-01-31T23:59:59Z). If omitted, defaults to the current time.
startTimestringThe start of the billing period to retrieve, specified as an ISO 8601 datetime string (e.g., 2023-01-01T00:00:00Z). If omitted, defaults to a reasonable historical starting point.

No output schema declared.

No examples provided.

list_network_volumes ~33

Retrieves a list of all network volumes available in the system. Use this operation to discover and enumerate network storage resources.

Input schema present but exposes no named parameters.

No output schema declared.

No examples provided.

list_pod_billing_history ~293

Retrieve aggregated billing records for your Pods over a specified time period. Results can be grouped by individual Pod or GPU type, with flexible time bucket granularity.

NameTypeReqDescription
bucketSizestringTime granularity for aggregating billing records. Choose from hourly, daily, weekly, monthly, or yearly buckets. Defaults to daily aggregation.
endTimestringEnd of the billing period to retrieve, specified as an ISO 8601 datetime (e.g., 2023-01-31T23:59:59Z). If omitted, defaults to the current time.
gpuTypeIdstringFilter results to Pods equipped with a specific GPU type. Accepts the full GPU model name (e.g., 'NVIDIA GeForce RTX 4090'). Omit to include all GPU types.
groupingstringOrganize billing records by Pod ID or GPU type. Defaults to grouping by GPU type. Use 'podId' to see per-Pod breakdowns.
podIdstringFilter results to a single Pod by its ID. Omit to include all Pods.
startTimestringStart of the billing period to retrieve, specified as an ISO 8601 datetime (e.g., 2023-01-01T00:00:00Z). If omitted, defaults to 30 days before the end time.

No output schema declared.

No examples provided.

list_pods ~403

Retrieve a list of Pods with optional filtering by compute type, hardware specifications, location, status, and other attributes. Supports inclusion of related metadata such as machine details, network volumes, savings plans, and templates.

NameTypeReqDescription
computeTypestringFilter results to only GPU-based or CPU-based Pods.
cpuFlavorIdarrayFilter to CPU Pods matching any of the specified CPU flavor identifiers (e.g., cpu3c, cpu5g).
dataCenterIdarrayFilter to Pods located in any of the specified RunPod data center regions (e.g., EU-RO-1).
desiredStatusstringFilter to Pods currently in a specific operational state: RUNNING, EXITED, or TERMINATED.
endpointIdstringFilter to worker Pods associated with a specific Serverless endpoint ID. Worker Pods are excluded from results by default unless includeWorkers is enabled.
gpuTypeIdarrayFilter to Pods with any of the specified GPU types attached (e.g., NVIDIA GeForce RTX 4090, NVIDIA RTX A5000).
idstringFilter to a specific Pod by its unique identifier.
imageNamestringFilter to Pods created with a specific container image.
includeMachinebooleanInclude detailed information about the physical machine or node the Pod is running on.
includeNetworkVolumebooleanInclude detailed information about any network volume attached to the Pod.
includeSavingsPlansbooleanInclude information about active savings plans or discounts applied to the Pod.
includeTemplatebooleanInclude information about the Pod template used during creation, if applicable.
includeWorkersbooleanInclude Serverless worker Pods in the results. By default, only standard Pods are returned.
namestringFilter to Pods with a specific name.
templateIdstringFilter to Pods created from a specific template ID.

No output schema declared.

No examples provided.

list_templates ~105

Retrieve available templates with optional filtering to include endpoint-bound, community, and official Runpod templates. By default, returns only your personal templates.

NameTypeReqDescription
includeEndpointBoundTemplatesbooleanInclude templates that are bound to Serverless endpoints in the response. Disabled by default.
includePublicTemplatesbooleanInclude community-made public templates in the response. Disabled by default.
includeRunpodTemplatesbooleanInclude official Runpod templates in the response. Disabled by default.

No output schema declared.

No examples provided.

reset_pod ~55

Reset a Pod to its initial state, clearing any runtime state or configuration changes. This operation restarts the Pod and restores it to a clean state.

NameTypeReqDescription
podIdstringyesThe unique identifier of the Pod to reset.

No output schema declared.

No examples provided.

restart_pod ~54

Restart a running Pod, causing it to stop and start again. This operation is useful for refreshing a Pod's state or recovering from transient issues.

NameTypeReqDescription
podIdstringyesThe unique identifier of the Pod to restart.

No output schema declared.

No examples provided.

start_pod ~49

Start or resume a Pod that is currently stopped or paused. This operation transitions the Pod to a running state.

NameTypeReqDescription
podIdstringyesThe unique identifier of the Pod to start or resume.

No output schema declared.

No examples provided.

stop_pod ~50

Stop a running Pod, halting its execution and resources. This operation gracefully terminates the Pod identified by the provided ID.

NameTypeReqDescription
podIdstringyesThe unique identifier of the Pod to stop.

No output schema declared.

No examples provided.

update_endpoint ~425

Modify configuration settings for a Serverless endpoint, including scaling behavior, resource allocation, worker limits, and attached volumes.

NameTypeReqDescription
cpuFlavorIdsarrayFor CPU endpoints, an ordered list of RunPod CPU flavor IDs to attach to workers. The list order determines rental priority.
dataCenterIdsarrayAn ordered list of RunPod data center IDs where workers can be deployed. Defaults to all available global data centers if not specified.
endpointIdstringyesThe unique identifier of the endpoint to update.
gpuCountintegerFor GPU endpoints, the number of GPUs to attach to each worker. Must be at least 1.
gpuTypeIdsarrayFor GPU endpoints, an ordered list of RunPod GPU type IDs to attach to workers. The list order determines rental priority.
namestringA user-friendly name for the endpoint. Names do not need to be unique and can be up to 191 characters.
networkVolumeIdsarrayA list of network volume IDs to attach to the endpoint, enabling multi-region storage access.
scalerTypestringThe autoscaling strategy: QUEUE_DELAY scales workers when requests exceed a latency threshold, while REQUEST_COUNT scales based on queue depth divided by a target ratio.
scalerValueintegerFor QUEUE_DELAY scaling, the maximum seconds a request can wait before triggering a new worker. For REQUEST_COUNT scaling, the target number of requests per worker. Must be at least 1.
templateIdstringThe template ID used to configure the endpoint's runtime environment and dependencies.
vcpuCountintegerFor CPU endpoints, the number of vCPUs allocated to each worker. Defaults to 2.
workersMaxintegerThe maximum number of workers that can run simultaneously. Must be 0 or greater.
workersMinintegerThe minimum number of workers that always run, even with no active requests. These are charged at a lower rate. Must be 0 or greater.

No output schema declared.

No examples provided.

update_endpoint_async ~483

Update configuration for a Serverless endpoint, including scaling behavior, resource allocation, and deployment settings. Changes apply to all future workers spawned on this endpoint.

NameTypeReqDescription
cpuFlavorIdsarrayFor CPU endpoints, an ordered list of RunPod CPU flavor IDs available for worker allocation. Earlier flavors in the list are prioritized for rental.
dataCenterIdsarrayAn ordered list of RunPod data center IDs where workers can be deployed. Earlier data centers are prioritized. Defaults to a global set of 26 data centers across multiple regions.
endpointIdstringyesThe unique identifier of the endpoint to update.
gpuCountintegerFor GPU endpoints, the number of GPUs to attach to each worker. Must be at least 1. Defaults to 1 GPU per worker.
gpuTypeIdsarrayFor GPU endpoints, an ordered list of RunPod GPU type IDs available for worker allocation. Earlier types in the list are prioritized for rental.
namestringA user-friendly name for the endpoint. Names do not need to be unique and can be up to 191 characters.
networkVolumeIdsarrayA list of network volume IDs to attach to the endpoint. Supports multiple volumes for multi-region deployments.
scalerTypestringThe autoscaling strategy: QUEUE_DELAY scales workers when requests exceed a latency threshold, while REQUEST_COUNT scales based on queue depth divided by a target ratio. Defaults to QUEUE_DELAY.
scalerValueintegerFor QUEUE_DELAY scaling, the maximum seconds a request can wait before triggering a new worker. For REQUEST_COUNT scaling, the target number of requests per worker. Must be at least 1. Defaults to 4.
templateIdstringThe template ID used to configure the endpoint's runtime environment and dependencies.
vcpuCountintegerFor CPU endpoints, the number of vCPUs allocated to each worker. Defaults to 2 vCPUs per worker.
workersMaxintegerThe maximum number of workers that can run simultaneously. Must be 0 or greater. Set to 0 for unlimited scaling.
workersMinintegerThe minimum number of workers that always run, even with no active requests. These workers are charged at a lower rate. Must be 0 or greater.

No output schema declared.

No examples provided.

update_network_volume ~118

Update the name and/or storage capacity of an existing network volume. Changes take effect immediately after the update is processed.

NameTypeReqDescription
namestringA user-defined name for the network volume. Names do not need to be unique and can be changed at any time.
networkVolumeIdstringyesThe unique identifier of the network volume to be updated.
sizeintegerThe new disk space allocation in gigabytes (GB) for the network volume. Must be greater than the current size and cannot exceed 4000 GB.

No output schema declared.

No examples provided.

update_network_volume_action ~112

Update the name and/or storage capacity of an existing network volume. The new size must be larger than the current allocated size.

NameTypeReqDescription
namestringA user-defined name for the network volume. Names do not need to be unique across volumes.
networkVolumeIdstringyesThe unique identifier of the network volume to update.
sizeintegerThe new disk space allocation in gigabytes. Must be between 0 and 4000 GB, and must exceed the current volume size.

No output schema declared.

No examples provided.

update_pod ~367

Update Pod configuration settings such as Docker image, environment variables, networking, and port mappings. Changes may trigger a Pod reset depending on the parameters modified.

NameTypeReqDescription
dockerEntrypointarrayOverride the Docker image's ENTRYPOINT instruction. Provide as an array of command segments (e.g., ['python', '-m', 'server']). An empty array uses the image's default ENTRYPOINT.
dockerStartCmdarrayOverride the Docker image's CMD instruction. Provide as an array of command segments (e.g., ['--port', '8080']). An empty array uses the image's default CMD.
envobjectEnvironment variables to set in the Pod runtime, provided as key-value pairs (e.g., {'ENV_VAR': 'value'}).
globalNetworkingbooleanEnable global networking for the Pod. Currently available only for On-Demand GPU Pods on select Secure Cloud data centers.
imageNamestringThe container image tag to run on the Pod (e.g., 'runpod/pytorch:2.1.0-py3.10-cuda11.8.0-devel-ubuntu22.04').
lockedbooleanLock the Pod to prevent stopping or resetting. Useful for protecting long-running workloads from accidental interruption.
namestringA user-defined name for the Pod. Names do not need to be unique and are limited to 191 characters.
podIdstringyesThe unique identifier of the Pod to update.
portsarrayList of ports to expose on the Pod. Each port is specified as 'port_number/protocol' where protocol is either 'http' or 'tcp' (e.g., ['8888/http', '22/tcp']).

No output schema declared.

No examples provided.

update_pod_request ~342

Update configuration settings for an existing Pod, including container image, environment variables, networking, and exposed ports.

NameTypeReqDescription
dockerEntrypointarrayOverride the Docker image's ENTRYPOINT instruction. Provide as an array of command arguments, or use an empty array to use the image's default ENTRYPOINT.
dockerStartCmdarrayOverride the Docker image's CMD instruction. Provide as an array of command arguments, or use an empty array to use the image's default CMD.
envobjectEnvironment variables to set in the Pod container, provided as key-value pairs (e.g., {'ENV_VAR': 'value'}).
globalNetworkingbooleanEnable global networking for the Pod. Currently supported only for On-Demand GPU Pods on select Secure Cloud data centers.
imageNamestringThe container image tag to run on the Pod (e.g., 'runpod/pytorch:2.1.0-py3.10-cuda11.8.0-devel-ubuntu22.04').
lockedbooleanLock the Pod to prevent stopping or resetting. Useful for protecting long-running workloads from accidental interruption.
namestringA user-defined name for the Pod. Names do not need to be unique and are limited to 191 characters.
podIdstringyesThe unique identifier of the Pod to update.
portsarrayList of ports to expose on the Pod. Each port is specified as 'port_number/protocol' where protocol is either 'http' or 'tcp' (e.g., ['8888/http', '22/tcp']).

No output schema declared.

No examples provided.

update_template ~336

Update an existing template's configuration, including Docker image settings, environment variables, exposed ports, and metadata. Changes apply to all Pods created from this template going forward.

NameTypeReqDescription
dockerEntrypointarrayOverride the Docker image's ENTRYPOINT instruction. Provide as an array of command segments (e.g., ['python', '-m', 'server']). Leave empty to use the ENTRYPOINT defined in the Dockerfile.
dockerStartCmdarrayOverride the Docker image's start command (CMD instruction). Provide as an array of command segments. Leave empty to use the CMD defined in the Dockerfile.
envobjectEnvironment variables to inject into Pods created from this template, specified as key-value pairs (e.g., ENV_VAR: value).
imageNamestringThe Docker image name and optional tag to use for Pods created from this template (e.g., 'myregistry/myimage:latest').
isPublicbooleanIf true, makes this Pod template visible to other Runpod users. If false, the template is private to your account.
namestringA human-readable name for the template.
portsarrayList of network ports to expose on Pods created from this template. Each port is specified as 'port_number/protocol' where protocol is either 'http' or 'tcp' (e.g., '8888/http', '22/tcp').
readmestringTemplate documentation in Markdown format, displayed to users when viewing or selecting this template.
templateIdstringyesThe unique identifier of the template to update.

No output schema declared.

No examples provided.

update_template_alternate ~293

Update an existing Pod template's configuration, including Docker image settings, environment variables, exposed ports, and metadata. Changes apply to all Pods created from this template going forward.

NameTypeReqDescription
dockerEntrypointarrayDocker ENTRYPOINT override for Pods using this template. Provide as an array of command segments; pass an empty array to use the ENTRYPOINT defined in the Dockerfile.
dockerStartCmdarrayDocker CMD override for Pods using this template. Provide as an array of command segments; pass an empty array to use the CMD defined in the Dockerfile.
envobjectEnvironment variables to inject into Pods using this template, specified as key-value pairs (e.g., ENV_VAR: value).
imageNamestringThe Docker image name and tag to use for Pods created from this template.
isPublicbooleanWhether this Pod template is visible to other Runpod users. Defaults to private (not visible).
namestringA human-readable name for the template.
portsarrayNetwork ports exposed by Pods using this template. Each port is specified as [port_number]/[protocol], where protocol is either 'http' or 'tcp'.
readmestringTemplate documentation in Markdown format, displayed to users viewing or using this template.
templateIdstringyesThe unique identifier of the template to update.

No output schema declared.

No examples provided.

Common questions

What is the com.mcparmory/runpod MCP server?

com.mcparmory/runpod is an MCP server listed in the public MCP registry as com.mcparmory/runpod. Launch, scale, and manage GPU pods and serverless endpoints across regions. This page covers its PyPI package (mcparmory-runpod).

Is the com.mcparmory/runpod MCP server safe to use?

com.mcparmory/runpod scores 60 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 20 September 2026. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

What tools does the com.mcparmory/runpod MCP server expose?

com.mcparmory/runpod exposes 35 tools: list_pods, create_pod, get_pod, update_pod, delete_pod, and 30 more. Their descriptions and schemas cost roughly 6,888 tokens of context every time the server is loaded.

Is the com.mcparmory/runpod MCP server still maintained?

com.mcparmory/runpod is still listed as active in the MCP registry. We last reached this channel on 20 September 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.