Retell AI MCP. Run full conversational agents from chat.
Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Retell AI connects your agent client to a full telecommunications stack. It lets you programmatically create, update, and manage voice agents, initiate outbound phone calls using `create_phone_call`, or register web audio interactions via text commands.
What your AI agents can do
Create agent
Creates a brand-new AI voice agent personality for your account.
Create llm
Sets up and configures the Large Language Model that drives the agent's conversation logic.
Create phone call
Starts an immediate, outbound phone call using specified source and target numbers.
Create, read, update, and list the unique AI personalities (agents) that handle calls.
Start an outbound phone call by providing source and destination numbers in a JSON payload.
Set up the underlying LLM that dictates how the agent behaves during a conversation.
Access detailed logs for specific calls, retrieving transcripts and sentiment analysis.
Register the necessary parameters to run a conversational agent within a web browser environment.
Ask AI about this MCP
Supported MCP Clients
Waiting for input…
Retell AI: 10 Tools for Conversational Agents
Use these ten tools to manage the entire lifecycle of your conversational bots. Create agents, initiate calls, and analyze call data directly through your chat client.
019d75fecreate agent
Creates a brand-new AI voice agent personality for your account.
019d75fecreate llm
Sets up and configures the Large Language Model that drives the agent's conversation logic.
019d75fecreate phone call
Starts an immediate, outbound phone call using specified source and target numbers.
019d75feget call details
Fetches a full record of a specific call, including the transcript and sentiment analysis.
019d75felist agents
Returns a list of every AI agent you currently have configured in your account.
019d75felist calls
Provides an overview and log of all historical and active calls processed by the system.
019d75felist phone numbers
Displays every phone number associated with your Retell account for use in calls.
019d75felist voices
Shows all the text-to-speech voices available to select for your agents.
019d75feregister web call
Sets up a new session that allows conversational AI interaction within a web browser context.
019d75feupdate agent
Modifies the instructions or parameters of an existing AI voice agent.
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 Retell AI, 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
This server hooks your agent client straight into a full telecom stack for building voice-driven agents. You ditch the Postman setup; you use simple chat instructions to manage the whole call life cycle—from setting up intelligent voices to placing actual calls.
Agent Management
Use create_agent when you wanna set up a brand new AI personality for your account. That function lets you define and create that unique voice bot agent. If you need to change what an existing agent knows or how it talks, run update_agent. This modifies the instructions or parameters of agents already running on your system.
To see every single AI agent you've configured, just call list_agents.
Conversational Configuration
To set up the brain for any agent, use create_llm. You configure the Large Language Model here; it dictates exactly how the agent talks and what its conversation logic is. The system needs to know which voices you wanna use—run list_voices to see every text-to-speech voice available for selection.
Initiating Calls and Web Interactions
Need to talk to somebody? You initiate an immediate, outbound phone call by calling create_phone_call, where you gotta provide both the source number and the target number in a JSON payload. If you need to know what numbers your account has available for these calls, run list_phone_numbers to see every associated number displayed.
For web interactions, use register_web_call. This sets up the parameters needed so your conversational AI can talk and listen right inside a web browser environment. You also have create_llm, which you're gonna need before any of these calls happen.
Monitoring Call History and Details
You gotta track what happens. To get an overview or log of every call, active or historical, you use list_calls. If you want the deep dive on a single interaction, run get_call_details. This function pulls a full record for one specific call, including transcripts and detailed sentiment analysis.
Grouping Capabilities: What You Can Do With These Tools
To build your agent's core identity, first use create_agent to establish the unique personality. Then you can modify that bot later using update_agent. If you need a full list of all agents configured across your account, check list_agents. The conversation logic is handled by setting up the underlying AI brain with create_llm.
If you're going to make phone calls, you start by checking available resources. Run list_phone_numbers to see which numbers are linked to your Retell account. You then kick off an immediate call using create_phone_call, specifying the source and destination numbers in the payload. For web-based conversations, calling register_web_call sets up the necessary audio session parameters within a browser context.
To keep tabs on what's been said, you query historical data using list_calls. If you need deep insights into one interaction, use get_call_details; this pulls transcripts and sentiment analysis for that specific call. Finally, if you wanna see which voices are ready to go, run list_voices.
How Retell AI MCP Works
- 1 First, ensure your AI client is authenticated with your Retell API Key. You'll need to install the Retell extension module in your MCP environment.
- 2 Next, tell your agent what you want using natural language—for instance: "Create a new support agent and run a test call."
- 3 The client executes the necessary tools (e.g.,
create_agent, thencreate_phone_call) sequentially, bringing back status reports or the conversation transcript.
The bottom line is: you talk to your chat agent like normal, and it translates those instructions into specific telecom API calls behind the scenes.
Who Is Retell AI MCP For?
This is for the platform architect or ops engineer who gets tired of stitching together multiple SaaS tools just to automate a simple call flow. If your business logic lives in natural language, but your action needs to ring a phone, this is what you use.
Uses the server to define and manage complex multi-step workflows: first calling list_agents to check existing bots, then using create_agent to spin up a new one for a pilot program.
Automates outbound communication flows by telling the agent to use create_phone_call with specific parameters whenever a customer needs follow-up support.
Manages the infrastructure by running diagnostics like list_calls or checking available lines using list_phone_numbers before deploying new features.
What Changes When You Connect
- Automate Calls: You don't need to write boilerplate code for outbound dialing. Just tell your agent to
create_phone_calland it handles the entire connection process, letting you skip manual API calls. - Deep Call Insight: After a call ends, you don't just get 'Success.' Use
get_call_detailsto instantly pull back the full transcript and know if the customer was positive or negative. - Agent Control: Need to tweak an agent's persona? Instead of recreating it, use
update_agent. You change its core instructions via chat, keeping your workflow simple. - Resource Visibility: Before you start a campaign, check what numbers and voices are available. The tools
list_phone_numbersandlist_voiceslet you validate resources before deployment. - Web Integration: Beyond phone calls, the
register_web_calltool lets your agent handle voice interactions directly inside a web browser session.
Real-World Use Cases
Post-Support Follow-up
A customer support manager needs to follow up after a ticket closes. Instead of manually dialing, they instruct their agent: "Call the client at +14085551234 using 'agent_support'." The agent uses create_phone_call and executes the script autonomously.
QA Bot Development
A development team needs to test a new AI personality. They run: "Create an agent called 'QADefault' and configure it using Model X." This uses create_agent followed by create_llm, allowing quick, iterative testing without touching code.
Daily Call Audit
An operations lead needs to check the mood of last week's sales calls. They simply ask: "Show me all call logs from Monday." The agent uses list_calls and then runs get_call_details on any specific ID provided.
Web Chat Upgrade
The team is upgrading their website chat to handle voice. They ask the agent: "Register a web call session for our new support bot." This runs register_web_call, making sure the audio pipeline works before going live.
The Tradeoffs
Assuming one tool handles everything
Trying to use only a messaging API when you actually need voice. You get flat, text-only transcripts and no call metadata.
→
You must combine tools. Start by checking resources with list_phone_numbers, then create the bot with create_agent, and finally run the test using create_phone_call.
Manually copying API payloads
Having to switch between Postman, documentation pages, and the chat interface just to set up a simple bot. It's slow and error-prone.
→ Keep it in the MCP chat. Just tell your agent: "Create an agent for sales using Model Y." The client handles the JSON payload for you.
Forgetting to check history
Running a new call without checking if there were previous conversations. This leads to repetitive scripts and poor customer experience.
→
Always start by asking: "What was the outcome of my last three calls?" This triggers list_calls and helps you contextualize your next action.
When It Fits, When It Doesn't
Use Retell AI if your core business process involves telecommunication—meaning, you need to initiate outbound phone calls or handle web-based voice interactions. If the goal is simply sending data records (like writing a CRM entry) or managing internal tickets, this server is overkill. You'll want it when 'talking on the phone' is part of the workflow. Don't use it if all you need is simple text messaging; in that case, stick to a dedicated chat API. However, if you are moving from basic text bots to actual voice agents, this is the standard toolset because it covers agent lifecycle (create_agent), conversation brain configuration (create_llm), and real-time communication (create_phone_call).
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Retell AI. 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
Having to switch between dashboards just to check call logs is a massive time sink.
Today, if you need to audit a conversation, you jump into the telephony dashboard. You find the call ID, then copy that ID and paste it into the analytics tool. If you want the transcript, you might have to click three different tabs—the metrics view, the recording link, and finally, the raw text log. It’s slow, and something always gets missed.
With Retell AI on your MCP server, you just ask: "What was the outcome of call ID c_f3a123?" The agent runs `get_call_details` immediately, giving you the transcript, sentiment analysis, and duration—all in one response. It’s that simple.
The Retell AI MCP Server: Voice Agent Management
Manually setting up a new bot means defining its personality, choosing the LLM, and then finding a test number to ring it out. This involves multiple logins, API keys in different places, and running scripts just to validate if 'Hello' works.
Here’s the fix: You tell your agent, "Build me a sales bot that calls this list of numbers." The client handles `create_agent`, configures the brain via `create_llm`, and then executes `create_phone_call`—all in one conversational turn. It makes building voice bots feel like chatting with a teammate.
Common Questions About Retell AI MCP
How do I list all my active AI agents using the Retell AI MCP Server? +
You use the list_agents tool. This command returns every configured bot ID and its basic status, letting you know what personalities are currently running in your account.
What information does `get_call_details` provide after a call? +
get_call_details gives you the full record: duration, sentiment score (e.g., 'positive'), and the complete, time-stamped transcript of what was said.
Can I make an outbound test call with `create_phone_call`? +
Yes. You run create_phone_call by passing a JSON payload that specifies both the source (from_number) and destination (to_number) you want to connect.
How do I update an existing bot's script? Do I use `update_agent`? +
That’s right. Use update_agent. You don't need to rebuild the whole agent; you just send a command that modifies its system instructions or personality parameters.
Before I start an agent, how do I check which voices are available using `list_voices`? +
It lists every text-to-speech voice your account supports. This tool helps you select the correct SSML parameters and voice IDs for your agents' personalities.
If an outbound call fails, how can I check the status using `list_calls`? +
You retrieve a list of all historical and active calls. Look for the specific call ID to see its final status (e.g., 'failed', 'disconnected') and associated error codes.
What is the role of `create_llm` when building an agent's conversational brain? +
It configures the LLM that drives the conversation logic for your agents. You use this tool to inject specific system instructions and choose the underlying language model.
How do I set up a web-based voice interaction using `register_web_call`? +
The command registers a unique session ID for your browser. This allows your AI client to manage real-time audio and interaction directly within the user's web environment.
Does Retell handle the telephone lines, or do I need my own Twilio/carrier? +
Retell AI is designed to integrate seamlessly into telephony workflows but generally focuses on the conversational agent side. They provide features to list and attach numbers (list_phone_numbers), but you often purchase those numbers or link a custom SIP trunk through Retell accounts. You can run list_phone_numbers to see your currently available direct lines.
How do I create an outbound call programmatically with my assistant? +
Simply ask your assistant to "Call X using my Y agent." Behind the scenes, the AI relies on the create_phone_call tool, transmitting a JSON payload supplying the from_number (a Retell registered phone number), to_number (destination), and an override_agent_id parameter to assign your voice agent correctly.
Can the AI initiate voice calls in a web browser instead of over the phone? +
Yes. Use the register_web_call tool to create a browser-based voice session. This returns a connection URL and access token that your frontend application can use to start a real-time WebRTC conversation with a Retell voice agent, without requiring any phone line.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Luma AI (Generative Video & Creative)
Generate cinematic AI videos and images via Luma — use Dream Machine for text-to-video, image-to-video, and professional camera control.
CodeRabbit
Manage AI-powered code reviews via CodeRabbit — list users, track PR review metrics, audit admin actions, and control seat assignments from any AI agent.
Poe
Manage AI chatbots on Poe — create bots, query other AI models, monitor messages, and track usage stats.
You might also like
Nozbe
Task management and team productivity.
Resend
Automate email delivery via Resend — send emails, manage domains, and track audiences directly from any AI agent.
Kit (ConvertKit)
Enable your AI agent to manage email subscribers, organize tags, and monitor broadcast campaigns via the Kit API.