Voiceflow MCP for AI. Test complex dialog flows without writing code.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
How this MCP server connects to your AI agent
Voiceflow MCP connects your conversational AI build directly into any agent client. This lets you design, test, and debug complex dialogue flows using a visual builder, without writing code or deploying new endpoints.
You can simulate user interactions, query the knowledge base for answers, and inspect every conversation transcript right from your chat interface.
What AI agents can do with Voiceflow Automation
Delete state
Resets the entire user conversation session variables.
Get feedback
Fetches the current upvote/downvote status for a project or interaction.
Get project
Retrieves detailed information about a specific Voiceflow project.
Send messages to your agent to instantly test conversational flows and expected reactions.
Ask the agent's connected knowledge base a question and get answers, or list the documents that power those answers.
Get, update, or completely reset the user's conversation variables to fix complex logic bugs.
List and fetch full transcripts for any project so you can audit how users actually talk to the bot.
Retrieve user feedback (upvotes/downvotes) and monitor project configurations live.
Ask an AI about this
Waiting for input…
What AI agents can do with Voiceflow: 12 Tools for Dialog Management
These tools allow you to manage every aspect of your conversation design process—from listing projects to checking user state.
Make your AI actually useful.
Add this MCP to Claude, Cursor, or Windsurf and your AI stops guessing. It gets real tools to look things up, take action, and handle the stuff you keep doing by hand.
Start using Voiceflow on VinkiusDelete State
Resets the entire user conversation session variables.
Get Feedback
Fetches the current upvote/downvote status for a project or interaction.
Get Project
Retrieves detailed information about a specific Voiceflow project.
Get State
Reads and returns the user's current conversation variables and state data.
Get Transcript
Gets the full details for one specific saved conversation transcript.
Interact
Sends a message to your agent, triggering the conversational flow logic.
List Kb Docs
Lists all the documents currently loaded into the knowledge base.
List Kb Tags
Shows available tags used for organizing and filtering KB documents.
List Projects
Retrieves a list of all existing Voiceflow projects under your account.
List Transcripts
Lists the IDs and details of stored user conversation transcripts.
Query Kb
Asks the knowledge base directly to find an answer based on provided context or...
Save State
Updates specific variables within the user's current conversation state.
Security and governance baked right in.
Pick your AI client below to get set up. Just create a Vinkius account, subscribe, and you're instantly up and running. We handle the entire backend infrastructure, delivering out-of-the-box support for HTTPS Streamable, SSE, and OAuth2—zero messy routing required.
Claude AI
Open Claude Settings
Go to claude.ai, click your profile icon, then navigate to Customize → Connectors.
Add Custom Connector
Click the "+" button and select Add custom connector. Paste your Vinkius endpoint URL:
https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp
Replace [YOUR_TOKEN_HERE] with your token
from cloud.vinkius.com. For OAuth-protected servers, expand
Advanced settings to add credentials.
Start a conversation
Open a new chat. The Voiceflow integration is available immediately — no restart needed.
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 Voiceflow, then connect any of our 5,100+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 5,100+ 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
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Voiceflow. 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
Built on the Model Context Protocol (MCP) for 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 connection provides 12 powerful capabilities that interface natively with Claude, ChatGPT, Cursor, and other compatible AI platforms. No middleware. No custom integration required.
Manual testing means endless clicking and copy/pasting.
Right now, if you want to test a complex user journey—say, an account upgrade that requires checking eligibility, updating the profile, and sending a confirmation email—you have to manually run through every single step. You click in one tab, copy data from another dashboard, paste it into a third tool, and then check the logs five times over just to make sure the conversation state didn't break anywhere.
With this MCP, you tell your agent client to start the flow. The agent handles all those steps behind the scenes using its tools. You simply watch the chat window; if something breaks, you immediately see the error and can use the session tools to fix it, saving hours of clicking.
Get a crystal-clear view into every conversation with Voiceflow.
Previously, checking user feedback meant opening separate analytics dashboards; verifying transcript details required digging through dated log files. You had to piece together the entire picture of success or failure by manually cross-referencing different data silos.
Now, you can pull all that information directly into your chat environment. Checking performance and auditing conversations becomes a single, simple request, letting you focus on building, not searching.
What your AI can actually do with this
This connector gives your AI agents an escape hatch. Instead of building massive backend services just to test if a chatbot works, you connect it directly through this MCP. You can send messages to simulate user conversations and immediately see how the agent responds across all its logic paths. Need to check what data the conversation is currently holding? It lets you retrieve, update, or even reset the user's entire session state for debugging.
If your agent uses a private knowledge base, you don't have to guess; you can query the documents and list exactly which files are powering the answers. Monitoring isn't just about seeing success messages. You get full visibility into what every AI agent is doing through Vinkius AI Analytics—you see the data flow and tool calls in real time.
This means when you build complex, multi-step workflows that chain together different services, you know exactly where the conversation broke down.
019dd184-6c3a-724f-861a-2e894ca64649 
Here's how it actually works
The bottom line is you manage your entire conversational ecosystem from one single AI client connection.
Subscribe to this MCP and provide your Voiceflow API Key and Version ID.
Your agent client connects, making the connector available for use in natural conversation prompts.
You ask your agent to perform an action, like listing all projects or querying a specific document, and it executes the task.
Who is this actually for?
Conversation Designers who spend hours clicking through flowcharts to test basic responses. AI Developers debugging complex state logic in a dev environment. Product Managers needing quick access to user feedback and conversation logs.
You use this MCP to quickly send messages and query the knowledge base, testing agent responses without touching the visual builder.
You manage user states by calling get_state or resetting variables with delete_state, letting you debug complex logic flows in real time.
You monitor the product's performance by retrieving user feedback via get_feedback and checking full transcripts using list_transcripts.
What Changes When You Connect
Debug agent logic instantly: Use get_state and save_state to read or write user variables, allowing you to fix state bugs without manual backend changes. This is critical when building complex multi-step processes.
Audit conversations thoroughly: Instead of relying on limited logs, you can list and fetch full conversation transcripts using list_transcripts and get_transcript, giving you a complete picture of user behavior for quality assurance.
Master your knowledge base: You don't just ask questions. Use list_kb_docs to see what data is available and then use query_kb to get answers, ensuring the agent pulls from the right source material every time.
Keep track of project status: Get an immediate view of overall performance by calling get_feedback, letting you know if users are struggling or loving a specific flow. This helps prioritize development efforts.
Inspect everything: The ability to list projects via list_projects lets you manage your entire conversational portfolio from one place, simplifying the operational oversight of multiple bots.
See it in action
The QA team needs to test a new refund path.
Instead of manually entering data into a staging environment, the agent client calls interact to start the conversation. The developer then uses get_state to check if the system correctly recorded the user's ID and purchase date throughout the dialogue.
A Product Manager wants to know why users drop off.
The PM asks the agent client to list all conversation transcripts. They review the results, identify a pattern of failure, and then use get_feedback to confirm if that pain point matches low user ratings.
A Developer is debugging complex logic.
The developer notices the agent is using an old variable. They call delete_state to wipe the current session, then use save_state immediately to inject the correct variables and test the fixed flow.
A Content team needs to verify policy answers.
The agent client asks the system about a new tax rule. The developer uses list_kb_docs first, confirming the source material is present, then relies on query_kb to ensure the answer reflects only the official document.
The honest tradeoffs
Testing flow logic by guessing states
The developer runs a few chats and just assumes variables are set correctly when they end. They don't check if the session variables persisted across calls.
Always use get_state after complex interactions to verify that every variable, like user role or order ID, was saved correctly before proceeding with subsequent logic.
Bypassing source verification
The agent gives an answer, and the developer assumes it's correct. They don't check if the knowledge base found supporting documentation.
When using query_kb, always verify that the system has listed relevant documents first, either by calling list_kb_tags or checking the output sources.
Over-relying on single tool calls
Trying to fix a bug just by sending one message (interact) without knowing what data was corrupted previously.
If something goes wrong, don't retry blindly. Use get_state first. If the state is bad, use delete_state and then try again.
When It Fits, When It Doesn't
Use this MCP if your primary bottleneck is testing or monitoring conversational logic. You need to see what data is flowing through the agent at every step—whether it's user input, session variables, or knowledge base sources. This connector excels when you are building complex automation that needs multiple services chained together; for example, chaining a messaging MCP with this one allows an agent to take feedback and update its own state. Don't use this if your only need is simple data retrieval from a single source (like just getting project names). For those cases, a dedicated list-only tool might be cleaner. But when the conversation itself is the product you're building, this MCP gives you total visibility.
Questions you might have
How do I use the Voiceflow MCP to debug my agent? +
Use get_state to read what variables the user currently has in session. If they are wrong, call delete_state and then use save_state to inject the correct values for retesting.
Can I list all available projects with Voiceflow? +
Yes, you can run list_projects. This shows you a roster of every bot or workflow you've created within your account in one simple command.
What is the difference between `query_kb` and `list_kb_docs`? +
list_kb_docs only gives you an inventory—it shows what documents are available. You use query_kb when you actually need to ask a question and get an answer based on those docs.
How do I check user satisfaction using Voiceflow? +
You call the get_feedback tool, which pulls current upvote or downvote data for your project. This lets you track sentiment without leaving your development environment.
How do I manage or reset user conversation context using `save_state`? +
You use save_state to update or retrieve variables, allowing your agent to maintain memory across multiple turns. If the flow gets confused, calling delete_state resets all session data back to zero.
What metadata can I get about a specific Voiceflow project using `get_project`? +
This tool pulls deep details on a selected project. You retrieve critical information like the project's version ID, its last modified date, and overall configuration status.
After calling `list_transcripts`, how do I fetch the full dialogue content for a specific session using `get_transcript`? +
First, use list_transcripts to get the session ID. Then, pass that ID into get_transcript to pull every single user input and agent response from the recorded conversation.
How do I check what tags are available for my knowledge base using `list_kb_tags`? +
This tool pulls a clean list of all custom tags applied across your KB documents. Checking these tags helps you organize and filter which specific sources your agent can reference during a query.
Can I query my Voiceflow Knowledge Base directly via AI? +
Yes! Use the query_kb tool with your question. Your agent will trigger the Voiceflow RAG system and return the answer based on your uploaded documents.
How do I see the transcripts for a specific project? +
Run the list_transcripts query with your Project ID. The agent will return a list of past conversation logs, which you can then inspect using get_transcript.
Is it possible to reset a user's session via AI? +
Absolutely. Use the delete_state tool and provide the User ID. This will permanently clear the conversation history and variables for that specific session.
We've already built the connector for Voiceflow. Just plug in your AI agents and start using Vinkius.
No hosting. No infrastructure. No complex setup.
All 12 tools are live and waiting.
You're up and running in seconds.
Gemini Vinkius gives your AI agents access to the full catalog of app connectors, all fully managed, secure, and enterprise-ready. One subscription, every tool you need.
Built, hosted, and secured by Vinkius. You just connect and go.