CrewAI Platform MCP. Control, monitor, and launch multi-agent workflows.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
CrewAI Platform MCP Server manages and executes complex, multi-agent workflows. It lets your AI client list existing agent workflows, kick off autonomous runs with custom JSON inputs, and monitor the entire execution lifecycle.
You can also check agent backstories, list all registered tasks, or instantly stop a runaway process using tools like `kickoff_crew` and `cancel_run`.
It's full orchestration control.
What your AI agents can do
Cancel run
Sends an interrupt signal to stop an active multi-agent run immediately.
Get agent
Retrieves the specific structured rules and properties defining an active agent.
Get crew
Extracts structural properties for a specific multi-agent workflow (crew).
Kick off a multi-agent process by providing a JSON payload, allowing your agent to run complex, pre-defined workflows.
Check the status of running workflows using get_run_status and stop them at any time with cancel_run.
List all available agents (list_agents) or workflows (list_crews) and retrieve their detailed structures.
Get detailed information on a specific agent's role (get_agent) or a task's definition (get_task).
Inspect the validation rules for asynchronous results and how the system handles data outside standard JSON formats (list_webhooks).
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
019d757ecancel run
Sends an interrupt signal to stop an active multi-agent run immediately.
019d757eget agent
Retrieves the specific structured rules and properties defining an active agent.
019d757eget crew
Extracts structural properties for a specific multi-agent workflow (crew).
019d757eget run status
Checks the current logging and status of an active agent execution run.
019d757eget task
Identifies the detailed parameters and status of a specific task within a workflow.
019d757ekickoff crew
Starts a new, complex multi-agent workflow by accepting a JSON payload.
019d757elist agents
Lists all deployed, defined agents within the CrewAI platform.
019d757elist crews
Identifies all defined, bounded multi-agent workflows (crews) in the system.
019d757elist tasks
Checks the history and status of all registered tasks in the system.
019d757elist webhooks
Retrieves information about configured webhooks and their validation criteria.
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 CrewAI Platform, 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
Yo, this CrewAI Platform MCP Server gives your AI client total control over complex, multi-agent workflows. You'll be running these things like a real command center, not some magic box. You can use list_agents to see every agent defined on the platform, and list_crews to check out all the pre-built multi-agent workflows.
You'll also find list_tasks lets you audit all the tasks registered in the system.
To kick off a whole new, complex workflow, you just use kickoff_crew, dropping in a JSON payload to start the action. You can monitor what's going down with get_run_status to see the current status and logs of an active run. If something goes sideways, you can immediately hit cancel_run to stop the process.
Need to dig into the details? You've got get_agent for the specific rules and properties of an agent, and get_task gives you the detailed parameters and status for any task inside a workflow. You can also check out the system's async data handling with list_webhooks to see the validation criteria for results that aren't standard JSON.
You'll even get the structural blueprints for a specific workflow using get_crew.
How CrewAI Platform MCP Works
- 1 Subscribe to the CrewAI Platform server and enter your unique CrewAI Agent Token (find this in the 'Status' tab of your Crew page in the AMP Dashboard).
- 2 Your AI client uses a natural language prompt (e.g., 'List all crews') to call the exposed tools.
- 3 The server executes the tool, returning structured data (e.g., a list of crew IDs or a run ID) that your AI client uses to continue the workflow.
The bottom line is, you talk to your AI client, and it talks to this server to run the complex agent logic for you.
Who Is CrewAI Platform MCP For?
This is for the AI Architect who needs to manage and verify complex, multi-agent graphs. It's for the Developer who needs to trigger and retrieve results from autonomous workflows directly in the IDE. Product teams use this to prototype agentic flows quickly, and DevOps teams use it to monitor and manage webhooks across environments.
Manages and verifies agent role-playing backstories and maps the topology of complex, multi-agent graphs.
Triggers autonomous workflows and retrieves task results directly from the development environment or chat interface.
Monitors active crew executions, manages webhook integrations, and controls the lifecycle of agent runs.
What Changes When You Connect
- You can launch complex, multi-step processes instantly. Instead of manually running scripts, just provide a payload to
kickoff_crewand the agents take over. - Never lose track of a running process. Use
get_run_statusto see exactly which agent is working and what step it's on, even if the task is running in the background. - Need to stop something?
cancel_runsends an immediate interrupt signal. You hard-stop a runaway agent process without diving into logs. - Verify agent roles and data flow. Use
list_agentsandget_agentto check the backstories and rules of every agent before you trust its output. - Manage dependencies and outputs.
list_webhookslets you see the exact validation rules for results that leave the standard JSON format. - Audit the whole graph.
list_crewsgives you a list of every defined workflow, andget_crewpulls the full structural blueprint for any of them.
Real-World Use Cases
Generating a market research report
The marketing analyst needs a report on AI trends. They ask their agent to run the 'Market-Research-Team' crew. The agent uses kickoff_crew with a JSON payload containing the topic. The agent then monitors the process using get_run_status until the final report is generated.
Debugging a failed automated pipeline
The devops engineer notices a workflow is failing. They use list_crews to find the relevant workflow, then get_crew to check its structure. They use list_tasks to see the history and pinpoint exactly which task caused the failure.
Checking an agent's credentials
Before using a new agent, the architect asks, 'What is agent X's role?' The agent calls get_agent, which returns the detailed role-playing backstory and rules, ensuring the agent has the right permissions.
Stopping a bad process
An autonomous job starts generating irrelevant content and needs to stop immediately. The agent uses cancel_run with the active run ID, hard-stopping the process and saving compute costs.
The Tradeoffs
Treating agents like single functions
Calling get_agent and assuming that agent can run a complex task. The agent only gets the agent's definition, not the execution ability itself.
→
To run the task, you must first use list_crews to find the workflow, then use kickoff_crew with the correct JSON input. This triggers the full, multi-agent process.
Forgetting to check status
Kicking off a crew using kickoff_crew and then immediately assuming the result is ready. The workflow runs asynchronously, and the result won't appear until the process is finished.
→
After calling kickoff_crew, always use get_run_status with the resulting run ID. Loop this call until the status changes from 'RUNNING' to 'COMPLETED'.
Ignoring task dependencies
Trying to check the status of a task using get_task before the necessary upstream agent has finished its work. The data will be incomplete or the tool will fail.
→
When It Fits, When It Doesn't
Use this if you need to manage, monitor, or trigger complex, multi-step agent processes. You need the full lifecycle control—from defining the graph (list_crews, get_crew) to kicking it off (kickoff_crew) and controlling it (get_run_status, cancel_run).
Don't use this if you just need to run a simple, single-step task. If a simple API call suffices, use a dedicated, single-purpose endpoint instead. Also, don't rely on list_tasks for current state; always use get_run_status for real-time execution tracking.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by CrewAI. 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 10 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Managing agent workflows used to mean clicking through five different dashboards.
Today, running a multi-step process means jumping between the Crew dashboard, the Task log, and the Webhook validation panel. You have to manually copy run IDs, switch tabs, and piece together the final status. It's slow, error-prone, and you're always guessing if the process actually finished or just timed out.
With the CrewAI Platform MCP Server, you just tell your agent to start the process. The server handles the whole graph. You get a single, reliable run ID and a clear status feed, letting you track the execution from start to finish without leaving the chat window.
Kickoff CrewAI Platform MCP Server: Launching autonomous workflows
Before this, starting a complex workflow meant manually configuring inputs and pressing 'run' in a dedicated UI. You had to worry about JSON structure and available credentials.
Now, your agent calls `kickoff_crew` with a simple JSON payload. The server immediately provisions the run, giving you a clean Run ID and the ability to monitor it via `get_run_status`. It's just that simple.
Common Questions About CrewAI Platform MCP
How do I start a multi-agent workflow using kickoff_crew? +
You start a workflow by calling kickoff_crew and passing a JSON payload that defines the initial input for the crew. The server responds with a new Run ID, which you then use to track the progress.
What is the difference between list_crews and get_crew? +
list_crews shows you a simple list of all available workflows. get_crew pulls the full, detailed structural blueprint and properties of a single, specific workflow.
How do I stop a running agent with cancel_run? +
Use the cancel_run tool and provide the active run ID. The server sends an immediate interrupt signal, hard-stopping the execution and preventing further compute charges.
Can I check what an agent is allowed to do with get_agent? +
Yes, get_agent retrieves the agent's explicit, structured role-playing rules and backstories. This lets you verify exactly what the agent is supposed to know and what actions it can take.
What does list_webhooks do? +
The list_webhooks tool shows you all configured webhooks. More importantly, it lets you inspect the exact validation criteria for async results, which is key for debugging data flow.
How do I use list_agents to find out which agents are available? +
list_agents lists all deployed agents. This tells you which role-playing agents are active in your environment and what capabilities they possess.
What does get_run_status tell me about an ongoing workflow? +
get_run_status retrieves the explicit cloud logging, showing the current step and state of any active crew run. You can track progress and pinpoint where a task might be stuck.
What kind of data can I get from list_tasks? +
list_tasks dispatches a validation check and routes the explicit gateway history for any given task. This helps you verify the execution history and expected inputs for modular operations.
Can my agent kickoff a new CrewAI workflow? +
Yes. Use the 'kickoff_crew' tool. Provide the Crew ID and a JSON object with the required inputs. The agent will activate the multi-agent processing immediately, returning a run ID for tracking.
How do I monitor the progress of an active agent run? +
Use the 'get_run_status' tool with your Crew ID and Run ID. Your agent will grab the live execution state, showing you which agents are currently working and which tasks have been completed.
Can I cancel a running crew via the agent? +
Absolutely. The 'cancel_run' tool dispatches an instant interrupt signal to the CrewAI platform, hard-stopping active LLM contexts and terminating the execution flow immediately.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
NVIDIA AI
Access LLMs, embeddings, code generation, and reasoning via NVIDIA API Catalog.
Home Assistant
Control smart home devices and automations via Home Assistant REST API — lights, climate, media, covers, sensors, and more.
Relevance AI
Equip your AI to trigger custom autonomous agents, execute chained prompts, and manage unstructured knowledge datasets directly within your Relevance AI studio.
You might also like
ORCID (Researcher IDs)
Connect to the ORCID registry to search for researchers, fetch biographical data, and manage research activities like works and funding.
Firecrawl
Crawl and scrape entire websites into clean LLM-ready markdown with a single API call that handles JavaScript rendering.
OpenMeter (Usage Metering for AI/Cloud Billing)
Meter AI usage, manage customer entitlements, and automate cloud billing via OpenMeter's high-performance usage tracking API.