# Conductor (Netflix OSS) MCP for AI Agents AI Agent Connect > Conductor (Netflix OSS) MCP. This connector lets you manage Netflix Conductor workflows, task definitions, and executions through natural conversation with your AI agent. It removes the need to jump between tabs or manual dashboards to handle your microservices orchestration. ## Overview - **Category:** industry-titans - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_TUjFjxN9H3g5Ba44H63GIIO4Av31lZmg9U3GcY2P/ai-agent-connect - **Tags:** workflow-orchestration, microservices, netflix-conductor, automation, process-management ## Description The Conductor (Netflix OSS) MCP connects your Netflix Conductor instance to your AI agent so you can handle complex microservices orchestration through a simple chat interface. Instead of hunting through a dashboard to find a specific workflow or manually clicking buttons to trigger executions, you can just ask your agent to do the work for you. You can pull up specific versions of your orchestration logic, create new task definitions, or start new instances without ever leaving your workspace. This setup helps you move faster by turning manual backend operations into quick commands. If you want to stop wrestling with the UI and start managing your distributed systems with plain language, this is the bridge you need. It is one of the many high-quality integrations available in the Vinkius catalog designed to make infrastructure management feel like a conversation. ## Tools ### bulk_pause Pause multiple workflows at the same time. Use this to stop a group of processes for maintenance. ### bulk_restart Restart a group of workflows simultaneously. Use this to reset multiple processes after a failure. ### bulk_resume Resume multiple paused workflows at once. This saves time when restarting a large batch of jobs. ### bulk_retry Retry a group of workflows that failed. This helps you quickly recover from a batch failure. ### bulk_search Search for workflows using their IDs in bulk. This makes it faster to find specific instances. ### bulk_terminate Terminate a large group of workflows at once. Use this for an emergency stop of multiple processes. ### create_event_handler Create a new event handler in your instance. This allows you to set up new automated responses. ### create_task_definitions Create new task definitions for your workers. This helps you set up the underlying logic for your tasks. ### create_workflow_definition Create a new workflow definition. Use this to add new orchestration logic to your system. ### delete_event_handler Delete an existing event handler. Use this to clean up old automated triggers. ### delete_task_definition Delete a specific task definition. Use this to remove old or unused task types. ### delete_workflow_definition Delete a workflow definition from your instance. This helps keep your system clean of old logic. ### execute_workflow Execute a workflow synchronously. Use this when you need the AI to wait for the result. ### get_all_queues Get the pending counts for all queues. This helps you see a high-level view of your system load. ### get_correlated_workflows Get workflows grouped by correlation ID. This is great for finding all steps in a single user request. ### resume_workflow Resume a paused workflow. Use this to start a process again after a maintenance break. ### retry_workflow Retry the last failed task in a workflow. This helps you recover from a single point of failure. ### add_task_log Add a log entry to a specific task. This helps you keep a record of what happened during a process. ### bulk_remove Remove multiple workflows from the system at once. This is useful for clearing out old test data. ### get_event_handlers Get a list of all event handlers. Use this to see what automated triggers are currently active. ### get_queue_size Get the queue depth for a specific task type. This helps you identify bottlenecks in your system. ### get_running_workflows Get the IDs of all workflows currently running. Use this to see what is active right now. ### get_task_definition Get the details of a task definition by name. Use this to check the config of a specific task. ### get_task_definitions Get all task definitions in your instance. This gives you a full list of your available task types. ### get_task_logs Retrieve the logs for a specific task. This is your go-to for debugging a failed step. ### get_workflow_definition Get a specific workflow definition by name. Use this to inspect the logic of a single workflow. ### get_workflow_definitions Get all workflow definitions in your system. This provides a full list of your orchestration logic. ### get_workflow_names_and_versions Get a list of all workflow names and versions. This helps you see what logic is currently deployed. ### get_workflow_tasks Get the tasks for a specific workflow execution. Use this to see every step in a running process. ### get_workflow Get the details of a workflow execution by its ID. Use this to check the status of a specific run. ### pause_workflow Pause a single workflow. Use this to stop a process without deleting it. ### poll_batch_tasks Long poll for the status of multiple tasks. This is useful for checking a group of jobs at once. ### poll_task Poll for the status of a single task. Use this to wait for a specific step to finish. ### remove_workflow Remove a workflow from the system. Use this to delete an execution instance. ### requeue_tasks Requeue pending tasks. Use this to put jobs back in line if they need to be processed again. ### rerun_workflow Rerun a workflow from a specific task. Use this to restart a failed process from a certain point. ### restart_workflow Restart a workflow from the very beginning. Use this to reset a process entirely. ### search_workflows Search for workflows and get a summary. This is a quick way to find what you need. ### search_workflows_v2 Search for workflows and get the full objects. Use this when you need deep details on your results. ### skip_task Skip a specific task in a running workflow. Use this to move past a step that is no longer needed. ### start_workflow Start a new workflow execution asynchronously. Use this to kick off a process and keep moving. ### terminate_workflow Terminate a running workflow. Use this to kill a process that is stuck or no longer needed. ### update_event_handler Update an existing event handler. Use this to change how your system responds to specific triggers. ### update_task_by_ref Update a task using its reference name. This helps you modify specific task logic dynamically. ### update_task_definition Update a specific task definition. Use this to change the configuration of a task type. ### update_task Update the result of a task. Use this to manually set the output of a step in your workflow. ### update_task_v2 Update a task and poll for the next available task. This is great for manual step-by-step progression. ### update_workflow_definitions Create or update multiple workflow definitions at once. Use this for rapid deployment of changes. ### validate_workflow_definition Validate a workflow definition without saving it. Use this to check your JSON for errors first. ## Prompt Examples **Prompt:** ``` Show me all the workflows we have right now. ``` **Response:** ``` I've pulled the list of workflows from your Conductor instance. You currently have: * **user_onboarding** (v1, v2) * **payment_processing** (v1) * **inventory_check** (v1, v2, v3) ``` **Prompt:** ``` Can you check the queue size for the 'email_sender' task type? ``` **Response:** ``` The queue for **email_sender** currently has **42** pending tasks. Would you like me to check the pending counts for all queues to see if there are other bottlenecks? ``` **Prompt:** ``` I need to pause all workflows that are currently running. ``` **Response:** ``` I've identified the running workflows. I'll go ahead and use the bulk pause command to stop them now. Let me know if you need me to resume any specific ones later. ``` ## Capabilities ### Trigger workflow executions Start new workflow instances immediately through your AI client. ### Manage definitions Create, update, and delete workflow and task definitions with ease. ### Perform batch actions Pause, restart, or terminate large groups of workflows in one go. ### Monitor queue health Check pending counts and queue depths for your task types. ### Validate JSON logic Check your workflow definitions for errors before you save them. ### Retrieve task logs Get detailed logs for specific tasks to help with debugging. ## Use Cases ### Emergency Workflow Halt A production workflow gets stuck in a loop. The engineer asks the agent to use `bulk_pause` on all workflows with a specific ID to stop the cycle immediately. ### Rapid Testing A developer needs to see if a new task works. They ask the agent to use `create_task_definitions` and then `execute_workflow` to run a test instance in a sandbox. ### Audit and Review An architect needs to see how a process is structured. They ask the agent to use `get_workflow_tasks` to see every step in a specific execution. ### Batch Updates A team needs to update 10 different workflows with a new timeout. They use `update_workflow_definitions` to do it in one go instead of clicking 10 times. ## Benefits - Stop clicking through nested menus by using `get_workflow_definitions` to see your entire orchestration list in one view. - Prevent production errors by using `validate_workflow_definition` to check your JSON logic before it ever goes live. - Handle emergency maintenance faster with `bulk_pause` and `bulk_terminate` to stop large groups of processes instantly. - Debug failed steps in seconds by pulling specific logs with `get_task_logs` or adding custom notes with `add_task_log`. - Deploy changes across your system faster by using `update_workflow_definitions` to push updates to multiple workflows at once. - Monitor your system health from your chat by checking `get_all_queues` and `get_queue_size` for real-time bottleneck detection. ## How It Works The bottom line is you get a natural language interface for your entire Netflix Conductor orchestration layer. 1. Subscribe to the Connector and provide your Conductor URL. 2. Connect your AI client to the Vinkius host. 3. Ask your agent to list workflows, start executions, or update task definitions. ## Frequently Asked Questions **What is the Conductor (Netflix OSS) MCP?** It is a way to connect your Netflix Conductor instance to your AI agent. This allows you to manage, trigger, and debug your microservices workflows using natural language instead of a web dashboard. **Can I use this to start workflows?** Yes. You can ask your agent to start new workflow executions either synchronously or asynchronously depending on your needs. **Does it work with my existing Conductor setup?** Yes, it works with any standard Conductor instance. You just need to provide your Conductor URL when you connect the Connector. **Can I manage multiple workflows at once?** Yes. The Connector includes bulk tools that let you pause, restart, remove, or retry large groups of workflows in a single command. **How do I validate my workflow JSON?** You can simply ask your agent to validate a workflow definition. It will check the JSON for errors without actually saving it to your system. **Can I see logs for specific tasks?** Yes. You can ask your agent to fetch logs for any specific task. This makes it much easier to debug failed steps without hunting through the UI. **Is this for production use?** Yes. It is designed to be a production-ready way for engineers to interact with their Conductor orchestration layer more efficiently. **Can I check if my workflow JSON is valid without saving it to the server?** Yes! Use the `validate_workflow_definition` tool. It allows you to submit a workflow definition JSON for validation, and the server will return any structural or logic errors without persisting the changes. **How do I see all available workflows and their versions?** You can use the `get_workflow_names_and_versions` tool to get a high-level list of all registered workflows and their associated version numbers. **Is it possible to inspect the configuration of a specific task type?** Yes, use the `get_task_definition` tool with the specific `taskType` name. This will return the complete task metadata, including retry logic, timeouts, and input/output keys.