Langflow MCP. Command, manage, and debug complex AI 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.
Langflow (Visual Multi-agent Orchestrator) is an MCP Server that lets you control complex, multi-step AI workflows through natural conversation. You can list, create, and manage entire project folders, run specific flows by ID or name, and even trigger external webhooks to start background jobs.
It's designed for AI engineers and product teams who need to treat complex AI logic like a system they can command directly from their agent client.
What your AI agents can do
Create flow
Creates a new, blank workflow definition within Langflow.
Create project
Sets up a new organizational folder for grouping related AI projects.
Create response
Acts as a general-purpose, OpenAI-compatible endpoint for generating responses based on a specified flow ID.
The agent can list, create, and update project folders to keep complex agentic workflows organized.
You can run defined AI flows using run_flow, supporting both text and chat inputs.
The agent can retrieve, update, or delete entire flows using tools like get_flow, update_flow, and delete_flow.
Start background processes or external systems by calling trigger_webhook in response to a natural language command.
Retrieve logs (get_logs), chat history (get_monitor_messages), and detailed execution traces (get_monitor_traces) for debugging.
The agent can download files (get_file_v2), list user files (list_files_v2), and manage project-related file assets.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
019e5d2ccreate flow
Creates a new, blank workflow definition within Langflow.
019e5d2ccreate project
Sets up a new organizational folder for grouping related AI projects.
019e5d2ccreate response
Acts as a general-purpose, OpenAI-compatible endpoint for generating responses based on a specified flow ID.
019e5d2cdelete file v2
Removes a specific file asset from the Langflow environment.
019e5d2cdelete flow
Permanently deletes an existing AI workflow definition.
019e5d2cdelete project
Removes an entire project folder and all its contained assets.
019e5d2cget file v2
Downloads a specific file asset by its ID or path.
019e5d2cget flow
Retrieves all details for a single, existing workflow definition using its unique ID.
019e5d2cget logs
Fetches the most recent operational logs from the Langflow platform.
019e5d2cget monitor messages
Retrieves the full chat history and conversation messages from a monitoring run.
019e5d2cget monitor traces
Gets detailed execution traces, including span trees and component interaction logs, for debugging complex runs.
019e5d2cget monitor transactions
Retrieves logs detailing how different components interacted during a workflow run.
019e5d2cget project
Get project details
019e5d2clist files v1
List files for a specific flow (v1)
019e5d2clist files v2
List user files (v2)
019e5d2clist flows
List all flows
019e5d2clist projects
List all projects
019e5d2clist users
List all users (requires superuser)
019e5d2crun flow
Supports chat or text inputs. Execute a Langflow flow
019e5d2crun workflow
Run a workflow (v2 API)
019e5d2ctrigger webhook
Trigger a Langflow webhook
019e5d2cupdate flow
Update an existing flow
019e5d2cupdate project
Update project info
019e5d2cwhoami
Get current authenticated user info
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 Langflow (Visual Multi-agent Orchestrator), 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
Langflow MCP Server - Manage Agent Workflows
This server lets your agent take total control of your Langflow environment, treating complex, multi-agent AI logic like a system it can command. You'll use natural conversation to list, create, and manage entire project folders, run specific flows, and even trigger external webhooks. It's built for AI engineers and product teams who need to treat complex AI pipelines like system calls.
Manage Projects and Folders
Your agent can organize your work by creating a new project using create_project, getting details on existing ones with get_project, and listing all projects available via list_projects. It can also update project info with update_project, and if you're done with a whole project, it'll delete it using delete_project. You can also manage the underlying files: it lets you list user files with list_files_v2, download specific assets with get_file_v2, and delete assets using delete_file_v2.
Execute and Control Flows
Your agent can run a specific flow using run_flow, supporting both chat and text inputs. You can manage the entire lifecycle of your flows; retrieve all details for an existing flow with get_flow, create a brand new, blank workflow definition using create_flow, update an existing flow with update_flow, and permanently delete a flow definition using delete_flow.
You can list all available flows with list_flows.
Debugging and Monitoring
When a flow runs, your agent can check what happened. It gets the most recent operational logs from get_logs, retrieves the full chat history and conversation messages from a run with get_monitor_messages, and gets detailed execution traces, including span trees and component interaction logs, for debugging complex runs using get_monitor_traces.
You can also see how different components interacted during a run by checking get_monitor_transactions. For file assets specific to a flow, it lists them with list_files_v1.
Advanced Operations
Your agent can start background processes or talk to external systems by triggering a webhook using trigger_webhook. It can also run a workflow using the v2 API with run_workflow, and generate responses from a specified flow ID using create_response. You can also check the current user's info with whoami.
Cleaning Up and Listing Users
It lets you list all users with list_users (this requires superuser permissions). It also lets you list files for a specific flow using list_files_v1.
How Langflow MCP Works
- 1 Subscribe to the server and input your Langflow Base URL and API Key into your AI client.
- 2 Direct your AI client to use a tool (e.g.,
list_flows) to check the status or availability of your workflows. - 3 Ask the agent to perform a complex action (e.g., 'Run the Market Analyzer flow'). The agent calls the
run_flowtool, and the results come back in the chat.
The bottom line is, you get to manage your entire visual AI environment—from creating projects to running flows and debugging—all through simple conversation with your agent.
Who Is Langflow MCP For?
This server is for AI Engineers, Product Managers, and DevOps Leads who need to test, manage, and integrate complex agentic workflows without switching contexts. If you're tired of clicking through a dashboard or running commands in a separate terminal just to check a flow's status, this is for you.
Testing and triggering complex RAG or multi-agent flows directly from the chat or code editor. They use run_flow to test inputs and get_monitor_traces to see where the logic failed.
Monitoring and organizing production assets. They use list_projects and list_flows to keep track of versioning and which workflows are live.
Integrating AI logic into larger systems. They use trigger_webhook to start background workflows when an external event happens, or run_workflow for scheduled tasks.
What Changes When You Connect
- Run any flow instantly. Use
run_flowto execute workflows by ID or name. You don't need to copy-paste inputs; just tell the agent to run it with the necessary data. - Debug complex logic. Need to know why a flow failed? Use
get_monitor_tracesto pull execution traces and span trees. It shows exactly which component interaction caused the issue. - Structure your work. Keep your AI assets clean. Use
list_projectsandcreate_projectto organize your flows into logical folders, making sure nothing gets lost. - Automate external triggers. Don't just run flows manually. Use
trigger_webhookto start a background workflow when an external system sends an event. - Maintain full visibility. Check the status of everything with
get_monitor_messagesto see the full chat history, orget_logsfor a quick overview of recent activity. - Manage assets easily. If you need a file or want to delete a flow, use
get_file_v2ordelete_flow. It keeps your development loop fast and contained.
Real-World Use Cases
Testing a new data pipeline
A data scientist builds a multi-step RAG pipeline. Instead of opening the Langflow UI, they ask their agent: 'Run the Data Extraction Pipeline with the company's latest report.' The agent calls run_flow, and the results are immediately available for review in the chat. They then use get_monitor_traces to verify the data sources.
Organizing production agents
A product team has dozens of experimental flows. They ask the agent to 'List all my project folders.' The agent responds with IDs, allowing the team to use create_project and delete_project to move the experimental flows into a 'V2 Production' project folder.
Responding to an external alert
A CI/CD system detects an anomaly and sends a webhook. The team doesn't need to log into Langflow. They simply call trigger_webhook via the agent, which kicks off an immediate, resource-intensive diagnostic workflow, and the results are streamed back.
Updating a failing agent
An agent flow breaks because a dependency changed. The engineer asks the agent to 'Get the details for the failing flow.' The agent calls get_flow to retrieve the configuration, which the engineer can then update_flow before re-running it with run_flow.
The Tradeoffs
Trying to manage everything manually
Opening the Langflow UI, navigating to the project list, then opening the flow, and finally running it—all while trying to copy-paste the input prompt somewhere else.
→
Use the agent. First, ask the agent to list_projects to find the right folder. Then, tell it to run_flow with the specific flow name and the input data. It keeps everything in one conversation.
Assuming tools know the data context
Asking the agent to 'run the flow' without specifying the correct flow ID or project folder, which leaves the agent guessing or failing with a generic error.
→
Always check the scope first. Use list_flows to see the exact names and IDs. If it's a large project, use list_projects to narrow down the context before calling run_flow.
Debugging by guesswork
Seeing a failed run and guessing which component failed, or manually trying to check logs across three different dashboards.
→
Use get_monitor_traces. It pulls the full execution path and component interaction logs (get_monitor_transactions) into one place. You don't guess; you read the evidence.
When It Fits, When It Doesn't
Use this server if your primary bottleneck is the friction between development and execution. If you need to manage complex, multi-step AI logic—like RAG or multi-agent systems—and you want to interact with that logic using natural language conversation instead of clicking through a GUI, this is for you. The tools like run_flow, get_monitor_traces, and list_projects give you direct, programmatic control.
Don't use this if you just need simple, isolated API calls (e.g., just saving a file). For that, a dedicated file management tool is better. Also, don't use it if your workflow doesn't involve state management; these tools are built for complex, stateful logic. If you only need to send a simple message, use a messaging service instead.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Langflow. 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 24 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Debugging a complex AI workflow shouldn't require jumping between five tabs.
When an agent fails, the manual process is a nightmare. You have to copy the run ID, jump into the Langflow UI, find the logs tab, then find the specific component that failed, and finally copy that error message into a ticket. It’s slow, and half the time you lose context.
With this MCP server, you just tell your agent, 'Show me the traces for the last run.' The agent calls `get_monitor_traces`, and the entire execution path, the component interactions, and the error message appear right in the chat window. It keeps the context right where it belongs.
Langflow MCP Server: Run and Manage Flows
Before this server, running a test flow meant opening the Langflow interface, selecting the flow by name, entering the input, and hitting 'Run.' If you wanted to change the input or re-run it, you had to repeat the whole process.
Now, you just tell your agent, 'Run the Market Analyzer flow with the prompt: [new data].' The agent calls `run_flow`, and the entire process—execution, completion, and results—is managed through a single command. It’s direct control.
Common Questions About Langflow MCP
How do I list all the workflows available in Langflow using the `list_flows` tool? +
You tell the agent to execute the list_flows tool. It returns a list of all defined flows and their unique IDs, so you know exactly what you can run next.
What if I want to run a flow that needs an external event to start? +
Use the trigger_webhook tool. This sends a signal to your Langflow instance, starting the workflow run without needing a manual click or input from your agent client.
Can I see the chat history of a specific flow run? +
Yes, use the get_monitor_messages tool. It retrieves the full chat history for that flow run, letting you review the conversation exactly as it happened.
Do I need superuser privileges to use the `list_users` tool? +
Yes, the list_users tool requires superuser permissions, so your agent client must be authenticated with those rights to use it.
How do I check the status of a project using the `get_project` tool? +
The get_project tool retrieves the full details for a given project ID. This includes project metadata, associated files, and the last known status of its contained workflows. You can use this to confirm if a project is active or needs updates.
What if I need to delete an entire flow or project using the `delete_flow` or `delete_project` tools? +
These tools permanently remove the specified resources. Before running them, always confirm the correct ID. Deleting a project removes all associated flows and files within that project folder.
How can I list all the files available in a specific flow using the `list_files_v2` tool? +
The list_files_v2 tool fetches all user files associated with the current scope. You provide the relevant flow or project ID, and it returns a list of file names and their paths, allowing you to download them using get_file_v2.
What information does the `get_monitor_traces` tool provide? +
This tool gives you the full execution history, showing interaction logs and span trees. It pinpoints exactly where a flow failed or slowed down, letting you diagnose component interactions step-by-step.
Can I run a flow using its name instead of a long UUID? +
Yes! The run_flow tool accepts either the Flow ID or the Flow Name in the flow_id parameter, making it easy to trigger specific logic by name.
How do I see all the available projects and folders in my Langflow instance? +
Use the list_projects tool. It will return a list of all projects (folders) which help organize your flows and components.
Is it possible to trigger a flow from an external webhook payload? +
Absolutely. Use the trigger_webhook tool by providing the flow_id and the data JSON payload you want to send to the flow's entry point.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Nationalize
Predict the nationality of a person based on their name using the Nationalize.io API.
Workflow Orchestrator Prover
AI agents build fragile pipelines that fail silently, ignore rate limits, and double-process events. This prover enforces distributed systems discipline: mandatory dead-letter queues (DLQ), exponential backoff, stateful idempotency tracking, and secure credentials.
LiteLLM (LLM Proxy & Spend Tracking)
Manage your LLM gateway via LiteLLM — generate API keys, track spending, and orchestrate model fallback paths.
You might also like
Grain Watch
Access silo temperature monitoring via Grain Watch — track grain temperature, humidity, hot spots, and spoilage risk from any AI agent.
Tencent Map LBS API
China's leading LBS platform — essential for WeChat Miniapps, providing precise mapping, routing, and POI data.
AI Receptionist
Let AI answer your business calls, book appointments, and handle customer inquiries around the clock without missing a beat.