# Create an action layer for AI agents with Dialpad MCP

Give your AI structured access to calls, messages, and meetings—so it can take action, not just analyze.

Dialpad MCP connects AI agents to real-time business communications—so they can take action across calls, messages, and meetings using structured tools instead of APIs.

### Turn real-time conversations into action

Most systems stop at insight. Dialpad turns conversations into action—so AI can follow through, not just analyze.

### Make AI actually do things

AI can reason. It still struggles to act. Connect your AI agents to real-time communications—so they can take action inside workflows.

### Bring your own tools

Reuse internal systems and workflows—no rebuild required.

### Operate with full context

Dialpad is the system where: conversation → signal → decision → action → learning

so every action happens with context.

### One system, not another layer

No stitching tools together. Your AI, your communications, and your workflows—connected in one place.

## Three ways to build with Dialpad MCP

APIs were built for developers. MCP is built for agents.

Best for Fast time-to-value

### Use Dialpad MCP (out-of-the-box actions)

Make your AI a Dialpad power user instantly.

- 100+ pre-built actions across calls, messages, meetings, transcripts, and settings
- No custom API work
- Works with MCP-compatible assistants and agents

```
$ mcp connect dialpad --key dp_••••••••
```

✓ Connected · 100+ tools available

```
>dialpad_list_recent_calls({ limit: 5 })
← 5 calls returned

>dialpad_get_transcript({ call_id: "cl_9f2b" })
← { sentiment: "positive",
action_items: ["Send pricing deck"] }

>dialpad_send_message({ to: "sarah@acme.com", body: "Following up..." })
← ✓ Message sent
```

Best for LLM & agent builders

### Connect your AI tools to Dialpad

Bring Dialpad into your existing AI workflows.

- Plug Dialpad into copilots, assistants, or agent frameworks
- Turn communication workflows into part of larger automations
- Move from insight → action inside your AI stack

```python
from dialpad_mcp import DialpadTools

from your_agent import Agent

agent = Agent(
    model="gpt-4o",
    tools=[ *DialpadTools.all(),  # 100+ tools
    *your_tools],  # keep what you have
)

result = agent.run(
    "Summarize Acme call + log to Salesforce"
)

# ✓ Transcript fetched · Action items logged
```

Best for Custom tooling

### Bring your own MCP tools

Extend Dialpad AI Agents with your own systems.

- Connect custom MCP servers
- Reuse internal tools, CRMs, ticketing systems, and workflows
- Orchestrate actions across Dialpad + your stack in one agent

```json
// mcp.config.json
{
    "mcpServers": {
        "dialpad": {
            "command": "dialpad-mcp",
            "env": { "API_KEY": "dp_••••••••" }
        },
        "salesforce": { "command": "sf-mcp" },
        "zendesk": { "command": "zendesk-mcp" }
    }
}
```

// Agent orchestrates all three in one workflow.

## FAQs

## Dialpad MCP

What is Dialpad MCP?

What is an MCP server, and how does Dialpad MCP work with AI agents?

What can AI Agents do with Dialpad MCP?

Can Dialpad MCP connect to my existing tools and workflows?