Argo Workflows MCP. Query, debug, and audit your entire K8s pipeline from chat.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Argo Workflows MCP Server automates Kubernetes orchestration. It lets your AI client query, list, and inspect the status of active, pending, or historical workflows, pods, and cron jobs across any connected Argo cluster.
You can deep-dive into resource trees to find exactly where a pipeline failed—all without touching `kubectl` or the web UI.
What your AI agents can do
Get server info
Gets general operational information about the Argo Workflows server.
Get workflow
Retrieves the detailed resource tree and status for a specific Argo workflow.
List archived workflows
Lists workflow records that have been archived in Argo's history for later review.
The agent lists all workflow executions in a namespace, whether they're running, waiting, or finished.
The agent drills down into a specific workflow instance to map its resource tree, node status, and pod parameters to pinpoint failure causes.
The agent lists parameterized WorkflowTemplates and scheduled CronWorkflows, letting you see what jobs are set to run and how they are structured.
The agent searches historical, archived workflow records to help trace patterns or diagnose issues that happened weeks ago.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
019d7552get server info
Gets general operational information about the Argo Workflows server.
019d7552get workflow
Retrieves the detailed resource tree and status for a specific Argo workflow.
019d7552list archived workflows
Lists workflow records that have been archived in Argo's history for later review.
019d7552list cron workflows
Lists all scheduled cron workflows within a specified namespace.
019d7552list workflow templates
Lists reusable, parameterized workflow templates available in a namespace.
019d7552list workflows
Lists all active workflow executions in a specified Kubernetes namespace.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Argo Workflows, then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
Argo Workflows MCP Server gives your AI client full visibility into your Kubernetes orchestration. You'll use it to query, list, and inspect the status of active, pending, or historical workflows, pods, and cron jobs across any connected Argo cluster. It lets your agent deep-dive into resource trees to find exactly where a pipeline failed, all without you touching kubectl or the web UI.
View all active and past workflows. Your agent can list every workflow execution in a specified namespace, letting you see if it's running, waiting, or finished. Inspect a workflow's failure point. You can drill down into a specific workflow instance to map its resource tree, node status, and pod parameters, pinpointing failure causes. Manage scheduled and reusable jobs. Your agent lists reusable WorkflowTemplates and scheduled CronWorkflows, so you can see what jobs are set to run and how they're structured. Audit past pipeline runs. You can search historical, archived workflow records to help you trace patterns or diagnose issues that happened weeks ago.
When you use the get_server_info tool, your agent pulls general operational information about the Argo Workflows server. You can check all active workflow executions by calling list_workflows in a specific namespace. To look at a specific workflow, your agent uses get_workflow to retrieve the detailed resource tree and status. If you wanna check what's scheduled, list_cron_workflows lists all scheduled cron workflows in a namespace, and list_workflow_templates lists the reusable templates available there.
You can also look through historical records using list_archived_workflows to see archived workflows.
How Argo Workflows MCP Works
- 1 Subscribe to the Argo Workflows server and provide your cluster's Server URL and RBAC Bearer Token.
- 2 Your AI client calls a tool (e.g.,
list_workflows) and passes the required namespace or workflow ID. - 3 The server executes the query against the Argo API and returns the structured data, which your agent then summarizes for you.
The bottom line is, you get a conversational API layer over your Argo cluster, letting you query complex orchestration data using plain language.
Who Is Argo Workflows MCP For?
DevOps engineers who hate spending hours in the Argo UI. Platform teams who need to debug complex, multi-stage pipelines and SREs who must monitor cluster health and historical job runs quickly. If your job involves diagnosing a failing deployment, this server is for you.
Debugging pipeline failures, checking node statuses, and auditing running jobs without leaving their primary terminal or chat window.
Monitoring complex ETL workflows and scheduled cron operations to ensure data pipelines run correctly and on time.
Quickly querying the overall health of the Argo server and retrieving historical metrics to predict and solve capacity issues.
What Changes When You Connect
- Check the status of any running job immediately. Use
list_workflowsto see all active and pending executions in a namespace without opening the dashboard. - Pinpoint failure causes instantly. Use
get_workflowto inspect the full resource tree of a failed pipeline, telling you exactly which node or pod crashed. - Manage scheduled jobs easily.
list_cron_workflowslets you see every recurring job defined in a namespace, so you don't have to check the cron manifest manually. - Understand reusable components.
list_workflow_templatesshows you all the parameterized templates available, so you know what components your team can build with. - Trace historical issues.
list_archived_workflowslets you search records of old runs, helping you find patterns or debug failures that happened months ago. - View server health. Use
get_server_infoto quickly confirm the connection and operational status of the Argo cluster.
Real-World Use Cases
Debugging a failed ETL pipeline
A data engineer's pipeline fails overnight. Instead of navigating to the Argo UI and clicking through dozens of tabs, they ask their agent: 'What is the deep resource tree for the failed job ID X?' The agent uses get_workflow and returns the exact node (e.g., 'S3 upload step') and the exit code, telling them it's a permission issue.
Auditing compliance for scheduled jobs
The SRE team needs to know every scheduled job running in the production namespace. They ask their agent to run list_cron_workflows. The agent returns a clean list, allowing the SRE to confirm the schedule and owners of every recurring job, without reading raw YAML files.
Checking a flaky development branch
A developer needs to see if a new workflow template is ready for testing. They ask the agent to run list_workflow_templates in the dev namespace. The agent shows them the available parameterized templates, confirming that the necessary components are defined and ready to be consumed.
Investigating a vague production slowdown
The ops engineer notices performance degrading and needs to see what was running recently. They ask the agent to use list_archived_workflows to pull reports from the last week, helping them identify resource spikes or unexpected historical job runs.
The Tradeoffs
Diving into the Argo Web UI
Logging into the web UI, navigating to the workflow list, filtering by date, then clicking into the job, and finally expanding the resource tree to find the error. This process takes 15+ clicks and is highly prone to human error.
→
Instead, ask your agent to run get_workflow with the job ID. It pulls the entire resource tree and failure reason directly into your chat window, bypassing the UI entirely.
Relying on `kubectl` CLI alone
Running complex, multi-line kubectl get commands that require knowing the exact resource type and namespace, which is tedious and error-prone.
→
Use the server's tools. Ask your agent to list_workflows for all jobs in the namespace. It handles the complex K8s plumbing and just gives you the readable list you need.
Confusing templates with active jobs
Assuming that seeing a template listed by list_workflow_templates means the job is running. Templates are just blueprints; they don't run themselves.
→
To check if a job is active, use list_workflows. To see the reusable blueprint, use list_workflow_templates.
When It Fits, When It Doesn't
Use this server if your primary job involves debugging, auditing, or listing Kubernetes workflows and jobs. You need to know the status of running pipelines, the definitions of reusable components, or the history of past runs.
Don't use this if you just need to deploy a simple, single-container job (use standard K8s deployment tools). Also, if you only need basic cluster health metrics (use a dedicated monitoring system). This server is for the complexity layer—the things that run after deployment, like data transformations and multi-step orchestration.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Argo Workflows. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 6 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Debugging a pipeline failure shouldn't feel like forensic accounting.
Today, when a multi-step data pipeline fails, you open the Argo Web UI. You have to click the workflow ID, then click the failed step, then find the pod name, and finally click the logs tab. You're copying names, switching tabs, and cross-referencing status codes. It's a mess.
With this MCP server, you just ask your agent: 'What went wrong with workflow X?' The agent runs the necessary tools, gathers the resource tree and node status, and gives you a clean summary right in the chat. You get the failure cause, not a screenshot of a cluster dashboard.
Argo Workflows MCP Server: Full Control Over Your Pipelines
You stop manually running `kubectl get` commands for status checks or listing templates across different namespaces. You stop refreshing the Argo UI every five minutes just to see if the job finished.
Your AI agent handles the complex API calls for you. It synthesizes the data from `list_workflows`, `list_cron_workflows`, and `list_workflow_templates` into actionable intelligence. It's pure operational visibility.
Common Questions About Argo Workflows MCP
How do I check if a workflow is currently running using the Argo Workflows MCP Server? +
Use list_workflows to see all active executions in a namespace. This tool returns a list of running, pending, or recently completed workflow IDs, allowing you to confirm the current status.
Can I check the deep status of a failed workflow using the Argo Workflows MCP Server? +
Yes. Call get_workflow and provide the workflow ID. This tool gives you the detailed resource tree and node status, pinpointing the exact component that failed and why.
What is the difference between `list_workflows` and `list_archived_workflows`? +
list_workflows shows currently active or recently completed jobs. list_archived_workflows searches historical records in the database, useful for debugging long-term patterns.
How do I see all scheduled jobs in my cluster? +
Run list_cron_workflows and specify the namespace. This tool targets all recurring, time-based jobs, which is a different resource from a manually triggered workflow.
How do I use `list_workflow_templates` to see reusable job patterns? +
This tool lists all parameterized WorkflowTemplates available in a namespace. You'll see components like model-training-tmpl or data-cleanup that you can adapt for new jobs. This saves you from redefining common job logic repeatedly.
What information does `get_server_info` provide about the Argo Workflows cluster? +
get_server_info provides basic operational details about the Argo Workflows server. It confirms connectivity and gives general health metrics, letting you quickly verify if the endpoint is up and accessible.
How do I check scheduled jobs using `list_cron_workflows`? +
list_cron_workflows shows all recurring CronWorkflows defined in a namespace. It lists the schedule frequency and the job that runs, helping you track scheduled tasks without manual checking.
What is the difference between `list_workflows` and `list_cron_workflows`? +
The list_workflows tool shows all active and recent workflow executions. list_cron_workflows, however, only lists the templates for scheduled jobs (CronWorkflows) that run on a set schedule.
Can my AI agent figure out exactly which pod/node failed in an active workflow execution? +
Yes. If a workflow fails, you can ask your agent to retrieve the workflow tree by name. The agent uses the get_workflow tool to inspect the deeply nested structure, traverse the active nodes, and pinpoint the exact step or container that resulted in an error state without you ever needing to click through the Argo UI.
Can I list only scheduled periodic jobs across my cluster? +
Absolutely. You can use the dedicated list_cron_workflows capability to isolate and return strictly workloads orchestrated on a time schedule across any namespace, saving you from parsing through thousands of isolated runs.
Do I need to expose my internal Kubernetes API to use this? +
No. The integration strictly interfaces with the Argo Server UI/API layer via standard REST traffic using a scoped ServiceAccount Bearer token. Your cluster's overarching master kube-apiserver remains safely isolated from external agentic logic.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Baserow
Manage no-code databases via Baserow — list tables, query rows, create records and manage schemas from any AI agent.
Datadog
Monitor applications via Datadog — query performance metrics, search logs, and list active monitors directly from any AI agent.
Cal.com
Control Cal.com infrastructure organically — book events, manipulate schedules, and track availabilities using an AI-native conversational flow.
You might also like
Zoho CRM Marketing
Manage marketing campaigns, track campaign leads and contacts — full marketing automation in Zoho CRM.
Zoho CRM Admin
Manage Zoho CRM users, roles, profiles, layouts, territories, and tags — complete admin control through conversation.
SAMHSA Treatment Locator
Find substance use and mental health treatment facilities across the US using official SAMHSA locator data.