MCP Tools and Tool Logs Documentation

This document provides a complete guide for configuring MCP integrations, connecting external platforms, managing AI agent tools, and monitoring tool executions using Tool Logs. The platform allows seamless integration with third-party services to enhance AI agent capabilities and automate workflows.

MCP Tools (Platform Integrations)

MCP (Model Context Protocol) Tools allow AI agents to integrate with external platforms such as Gmail, Zoom, Razorpay, and many other business applications. These integrations help agents perform platform-specific tasks directly from conversations.

Key Benefits of MCP Tools:

Connect AI agents with third-party platforms
Automate cross-platform workflows
Enable platform-specific integrations
Extend agent functionality with platform-specific tools
Reduce manual operations using automation

1.1 Platform Catalog

Navigate to Tools → MCP Tools → Platform Catalog to browse all available platforms. The catalog shows 28 platforms including Gmail, Zoom, Razorpay, Resend, IRCTC, Zoho CRM, Zenoti, and more. Each platform card shows its auth type, category tags, number of available tools, and a Connect button.

Figure 1 — MCP Tools: Platform Catalog. Browse and connect to 28 platforms including Resend, Gmail, Zoom, Razorpay, IRCTC, Zoho CRM, and Zenoti.

feature	desc
Search	Search platforms by name from the search bar.
Category filter	Filter by category — Email, CRM, Payments, Video, etc.
Status filter	View Active, New, Beta, or Coming Soon platforms.
Sort	Sort by Popularity or A–Z.
Grid / List view	Toggle between grid cards and list view.

1.2 Connecting to a Platform (Gmail Example)

The platform allows easy integration with supported MCP providers using OAuth credentials or API-based authentication. Steps to connect Gmail:

Navigate to MCP Tools and find the required platform.
Click Connect for the required platform.
Enter or Test Configuration to validate the connection.
Once connected, open the My Connections tab.
Use Show Tools to view all available tools for the selected platform.
Configure the discovered tools and make them available for AI agents.

Figure 2 — Gmail MCP OAuth Configuration. Enter Client ID and Client Secret, then click Authorize with Gmail to complete the OAuth 2.0 setup.

1.3 My Connections Dashboard

After connecting platforms, navigate to the My Connections tab to manage all active integrations. This dashboard lists every connected platform with its current status, available tools, and last tested timestamp.

Figure 3 — My Connections Dashboard. Shows all connected platforms — Gmail (Active, 1 tool) and OCR (Active, 3 tools) with Test and Disconnect actions.

col	desc
Platform	Name and label of the connected platform.
Status	Active or Inactive — shows if the connection is live.
Tools	Number of tools available from this connection.
Last Tested	Timestamp of the most recent test.
Connected	Date the connection was first established.
Actions	Show Tools, Test connection, or Disconnect.

1.4 Available Tools for a Connection

Click Show Tools on any connection row to expand and view all tools available from that platform. Each tool has a unique ID and a Configure button to adjust its settings before attaching it to an agent.

Figure 4 — Available Tools for OCR Connection. Click Show Tools to expand — Document QA, Extract Image, and Extract PDF are available to configure.

tool	id	desc
Document QA	ocr.document_qa_tool	Ask questions about document content using AI.
Extract Image	ocr.extract_image_tool	Extract and analyse image content from documents.
Extract PDF	ocr.extract_pdf_tool	Extract text and structured data from PDF files.

1.5 External Server Integration

The External Server feature allows secure integration with custom MCP servers or external CRM systems using HTTPS endpoints and Bearer Token authentication.

Enter a valid HTTPS Server URL
Provide Bearer Token credentials
Assign a Connection Label for easy identification
Click Test & Discover Tools to fetch available tools
Configure discovered tools before attaching them to agents

Figure 5 — External MCP Server Configuration. Enter a valid HTTPS Server URL, Bearer Token, and an optional Connection Label, then click Test & Discover Tools.

field	details
Server URL *	Must start with https://. Your MCP server must be accessible via HTTPS.
Bearer Token *	Sent as Authorization: Bearer <token> header for secure authentication.
Connection Label	Optional. A friendly name to identify this connection (e.g. My Company CRM).

Connecting Tools to an Agent

After configuring MCP or Custom Tools, they can be attached to AI agents from the Agent Configuration panel. This enables agents to use selected tools during conversations and automate platform-specific operations.

Navigate to Create Agent or Edit Agent.
Go to the Tools & Events tab under Agent Configuration.
Select MCP Platform Tools or Custom Tools tab.
Browse all available tools.
Select or deselect tools according to agent requirements.
Configure tool-specific settings if required.
Click Update Agent to save changes.

2.1 MCP Platform Tools

In the Agent Configuration → Tools & Events tab, click MCP Platform Tools to view all tools from your connected platforms. Check the tools you want the agent to use — selected tools show a green Selected badge. Use the gear icon to configure tool-specific settings. Click Update Agent to save.

Figure 6 — Agent Configuration › Tools & Events: MCP Platform Tools tab. Document QA and Extract PDF are selected (shown in green). Click Update Agent to save.

Figure 7 — MCP Tool Selection Interface. Check tools to enable them for your agent. Selected tools show a green Selected badge and a Configure gear icon.

2.2 Custom Tools Configuration

Switch to the Custom Tools tab to view and select your webhook and API tools. The Agent Tools & Capabilities summary shows:

System Tools — always active, provided by the platform.
Custom Tools — user-selected webhook or API tools from the Tools section.
Total Available to LLM — combined count of active tools the agent can use.

Figure 8 — Custom Tools Configuration. The Custom Tools tab shows Agent Tools & Capabilities — System Tools (always active) and user-selected Custom Tools (webhook / API).

Tool Logs

The Tool Logs section provides complete visibility into all tool executions performed by AI agents. It helps administrators monitor executions, debug issues, track failures, and analyse platform activity.

Tool Logs provide the following insights:

Tool executions and current state
Daily execution statistics
Success and failure rates
Execution timestamps
User inputs that triggered tools
Tool Arguments passed during execution
Tool Results returned by integrations
Filtering logs by Channel, Status, and Tool
Exporting logs in JSON format

Figure 9 — Tool Logs Dashboard. Shows Total Executions, Today, Success Rate, and Failed counts with per-entry detail: User Input, Tool Arguments, and Tool Result.

Each log entry shows:

Active Agent — which agent performed the tool call.
User Input — the message or trigger that initiated the tool.
Tool Arguments — exact parameters passed to the tool (expandable).
Tool Result — response data returned by the integration (expandable).
Status — Success or Failed with timestamp.

Best Practices

Use secure HTTPS endpoints for all integrations.
Avoid exposing sensitive API keys and tokens.
Regularly monitor Tool Logs for failed executions.
Enable only required tools for AI agents.
Test integrations before production deployment.

Conclusion

MCP Tools and Tool Logs provide a powerful ecosystem for extending AI agent capabilities through external integrations, automation workflows, and monitoring systems. By combining platform connections, custom servers, agent configuration, and execution logs, organizations can build reliable, intelligent, and automated AI solutions.