# Froged MCP

> Froged connects customer success, support chat logs, and behavioral events to your AI agent. It gives you a single source of truth for every customer interaction—from initial contact details to custom in-app activity tracking.

## Overview
- **Category:** customer-support
- **Price:** Free
- **Tags:** omnichannel-support, customer-retention, behavioral-tracking, ai-agents, customer-success, engagement-analytics

## Description

When an agent needs context, they don't want to jump between five different tabs. This MCP connects Froged’s entire platform directly to your AI client, giving it the data it needs right when you need it. Your agent can instantly check a customer's complete profile, see every chat conversation history across all channels, and even track custom events—like when they upgraded their plan. If Marketing runs an in-app campaign, your agent knows about it. You manage everything from one place. Since this MCP is hosted on Vinkius, you connect once to the catalog and get access to Froged's full suite of customer data tools through simple conversation prompts. It’s all about getting a unified view so you can reply accurately every single time.

## Tools

### verify_api_status
Checks the current connection status and health of the API link.

### get_contact_details
Fetches detailed metadata about a specific customer contact.

### get_chat_details
Retrieves the full conversation history for a specified support chat.

### list_marketing_campaigns
Provides a list of all active marketing campaigns running in the platform.

### list_cs_contacts
Retrieves a list of all contacts managed within the system.

### list_support_conversations
Lists multiple ongoing support chats to give an overview of current activity.

### list_behavioral_events
Returns a list of tracked user behavioral events for analysis.

### list_kb_articles
Fetches titles and summaries for published help articles in your knowledge base.

### send_chat_message
Sends a reply message directly into an active support conversation thread.

### track_custom_event
Records and logs a specific custom action or behavior performed by the user.

### upsert_contact
Creates a new contact profile or updates an existing one with new information.

## Prompt Examples

**Prompt:** 
```
List my 5 most recent active support conversations.
```

**Response:** 
```
Retrieving conversations... I found 5 active chats, including a pricing inquiry from 'jane@example.com' and a bug report from 'john@example.com'.
```

**Prompt:** 
```
Track the event 'plan_upgraded' for user 'customer@email.com'.
```

**Response:** 
```
Event tracked! The 'plan_upgraded' event has been successfully logged for customer@email.com. Any associated onboarding campaigns will now trigger.
```

**Prompt:** 
```
Show me the contact profile for 'jane@example.com'.
```

**Response:** 
```
Searching contacts... I found Jane Doe (jane@example.com). She signed up 30 days ago and last interacted with support yesterday regarding 'API limits'.
```

## Capabilities

### Build complete customer profiles
The agent retrieves or updates contact records with detailed metadata to create a 360-degree view of the user.

### Manage support communication
You can list active chat conversations across all channels and send direct replies to keep the dialogue going.

### Log customer actions in real-time
The agent tracks custom user behavior, logging specific events that trigger downstream marketing or success workflows.

### View support documentation
It fetches published help articles from the knowledge base so agents can quickly reference self-service material.

### Review active campaigns and events
You get lists of current marketing campaigns or recent behavioral events to understand user context.

## Use Cases

### A support agent needs context for a vague query.
The agent asks the AI client: 'What's wrong with customer X?' The agent immediately uses `get_contact_details` and then calls `list_support_conversations`. It surfaces not only the last chat transcript but also that the user recently triggered the 'plan_upgraded' event, giving the agent the full story before sending a reply via `send_chat_message`.

### Marketing needs to confirm an upgrade.
A CSM finishes a call and knows the customer just signed up for premium. Instead of logging into marketing, they prompt the AI client: 'Log plan upgraded event for user Z.' The system calls `track_custom_event`, ensuring the automated onboarding campaign starts instantly.

### A team needs to audit a contact record.
The manager asks the agent to summarize data for a specific client. The agent runs `get_contact_details` and then calls `list_behavioral_events`, showing the manager every time that user logged in, what features they viewed, and when.

### Responding to an old or inactive chat.
A support agent reviews a list of chats. They use `list_support_conversations` to find the thread ID, then call `get_chat_details` for full context. Finally, they reply using `send_chat_message`, keeping the whole interaction logged.

## Benefits

- Stop manually checking multiple dashboards. You can check a customer's entire history—from recent chats to their profile details—all by asking your agent one question.
- Keep marketing campaigns running smoothly. By using `track_custom_event`, you ensure that when a user performs an action, the right automated campaign triggers immediately.
- Respond faster and more accurately. Instead of searching through help docs, ask your agent to pull information from the knowledge base via `list_kb_articles` and use it in the reply.
- Maintain perfect data records. When contact details change, you can run `upsert_contact` directly through the MCP, keeping everything current without manual form filling.
- Get a full operational picture. You don't have to guess what's happening; your agent lists all active chats (`list_support_conversations`) so you know exactly where to focus.

## How It Works

The bottom line is, you treat the entire customer service stack—contacts, chats, and events—like a single database accessible through chat.

1. Subscribe to this MCP on the Vinkius Marketplace.
2. Enter your Froged API Key in your AI client's settings.
3. Start interacting with customer data using natural language prompts in your agent.

## Frequently Asked Questions

**How do I get an API Key for Froged?**
Log in to your Froged Admin Panel, navigate to Settings > Integrations > API, and you can generate or copy your active API Key.

**Can I track custom events to trigger campaigns?**
Yes! Use the 'track_custom_event' tool. Provide the event name and the user's email to log the behavior and trigger any associated workflows.

**Is it possible to reply to support chats through the agent?**
Yes, use the 'send_chat_message' tool with the specific Conversation ID to post replies directly to the omnichannel inbox.

**How do I sync a new user to Froged?**
Use the 'upsert_contact' tool. Provide the user's email address, and if they don't exist, a new profile will be created. If they do, their metadata will update.

**How do I get detailed contact metadata using the `get_contact_details` tool?**
Yes, this tool pulls a full profile of any customer. You'll retrieve more than just their name; you get detailed information like when they signed up and how many times they've interacted with support.

**Can I pull the entire conversation history using the `get_chat_details` tool?**
Absolutely. The `get_chat_details` tool retrieves the full thread of a support chat, not just the latest message. This gives your agent the complete context needed to handle follow-up questions.

**What is the function of the `list_kb_articles` tool?**
The `list_kb_articles` tool fetches all published help articles from your Knowledge Base. This lets your agent provide self-service links or quickly pull information to answer common customer questions.

**How do I list and manage multiple customers using the `list_cs_contacts` tool?**
The `list_cs_contacts` tool provides a comprehensive roster of all Froged contacts. You can quickly see who is in your system without needing to know their specific email address first.