Model Context Protocol (MCP)

● Live Production Server

Connect AI Agents to Live Business Call Intelligence

Empower Claude, Cursor, Windsurf, and custom AI agents to query bilingual call transcripts, analyze customer objections, fetch 0–100 quality scores, and review agent metrics via the open Model Context Protocol.

Connect via MCP Setup Guide Start 14-Day Free Trial

Scoped API key authentication · Enterprise tenant isolation · TLS 1.3

Instant Integration

Connect in 60 Seconds

Add the Nodzo MCP server endpoint to your AI client configuration file to start querying your team's calls directly.

web.nodzo.ai/mcp

~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows)

{
  "mcpServers": {
    "nodzo": {
      "url": "https://web.nodzo.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_NODZO_API_KEY"
      }
    }
  }
}

Replace YOUR_NODZO_API_KEY with your token from your Nodzo dashboard.Protocol: MCP v1.0.0

Server Capabilities

Tools & Resources Exposed by Nodzo MCP

When connected, your AI client can autonomously discover and execute these tools with structured inputs and typed JSON outputs.

nodzo_search_calls

Tool

Search company calling history using multi-parameter filters including agent, telephone number, sentiment, call outcome, and date ranges.

Parameters

query(string)Optional keyword or phrase to search within conversation transcripts.

agent_id(string)Filter by specific sales or support calling rep identifier.

customer_phone(string)Filter by client or customer telephone number.

direction(enum)'incoming' | 'outgoing' | 'all'

min_score(number)Minimum 0–100 AI call quality score threshold.

limit(number)Maximum results to return (default: 20, max: 100).

Example AI Query

“Search all calls from this week with quality score below 75”

nodzo_get_transcript

Tool / Resource

Retrieve the full dual-channel transcript of any recorded business call with precise speaker diarization (Agent vs. Customer) and timestamps. Fully supports Hindi, English, and Hinglish.

Parameters

call_id*(string)Unique identifier of the call record.

include_timestamps(boolean)Whether to return millisecond-level word timestamps (default: true).

format(enum)'json' | 'markdown' | 'dialogue_text'

Example AI Query

“Fetch the verbatim bilingual transcript of customer call #8410”

nodzo_get_call_analysis

Tool

Fetch structured AI intelligence for a call: 0–100 quality index, customer intent classification, sentiment breakdown, customer objections, and extracted follow-up commitments.

Parameters

call_id*(string)Unique identifier of the call record.

include_coaching(boolean)Include prescriptive AI coaching suggestions for the rep.

Example AI Query

“Inspect pricing objections and promised next actions on call #8410”

nodzo_get_team_metrics

Tool

Pull aggregated calling and performance analytics across teams or individual agents over a selected timeframe.

Parameters

team_id(string)Optional team or department identifier.

timeframe(enum)'today' | '7d' | '30d' | 'custom'

metrics(array)['total_calls', 'connected_rate', 'avg_duration', 'talk_time', 'top_objections']

Example AI Query

“Compare telecalling connected rates and average duration between Teams A and B”

nodzo_crm_sync_status

Tool

Verify live synchronization state between Nodzo calls and connected CRM systems (HubSpot, Zoho CRM, or custom webhooks).

Parameters

call_id*(string)Unique identifier of the call record.

crm_provider(enum)'hubspot' | 'zoho' | 'webhook'

Example AI Query

“Check if the follow-up note and transcript synced to HubSpot deal #512”

Agent Interactions

What You Can Ask Your AI via Nodzo MCP

Real prompt workflows used by cross-functional teams connecting Claude or Cursor to their sales call data.

Sales Leaders & Directors

AI Agent Workflow

Instant 1:1 Coaching Preparation

"Query Nodzo for rep Rahul's calls this week where call score was below 75. Summarize the recurring customer objections he faced and give me 3 specific coaching points for our 1:1 meeting."

Saves hours of manual call auditing before coaching reps.

Product & Marketing Teams

AI Agent Workflow

Direct Voice-of-Customer Extraction

"Search all customer conversations from the last 14 days mentioning our competitor or pricing. Group customer concerns into a structured product feedback matrix with verbatim quotes."

Bridges the gap between sales conversations and product roadmap.

Support & QA Managers

AI Agent Workflow

Escalation & Compliance Audits

"Scan recent support calls for negative sentiment spikes or unresolved complaints. Identify any calls where mandatory disclaimer phrases were skipped."

Guarantees 100% QA coverage through natural language AI queries.

Developers & AI Engineers

AI Agent Workflow

Autonomous Post-Call Agent Workflows

"Whenever a high-priority call completes, retrieve the AI summary via Nodzo MCP, draft a personalized follow-up proposal, and post a Slack alert for the account owner."

Automates end-to-end post-call administration.

Enterprise Security & Privacy Architecture

Built to satisfy corporate data governance and compliance guidelines.

Scoped Token Control

Generate scoped API keys with fine-grained tool permissions. Revoke or rotate keys instantly from the manager portal.

Zero Public Model Training

Your business phone recordings and customer transcripts are never used to train public LLMs or sold to third parties.

Tenant Data Isolation

Queries are strictly locked to your authorized organization tenant. Cross-tenant leakage is architecturally prohibited.

45-Day Audio Alignment

Transcripts and analysis follow your company data policy, while raw audio files are automatically purged after 45 days.

Knowledge Base

Frequently Asked Questions

Technical and operational answers for teams integrating with Nodzo MCP.

What is Model Context Protocol (MCP)?

Model Context Protocol (MCP) is an open standard developed by Anthropic that allows AI applications (such as Claude Desktop, Cursor IDE, Windsurf, or custom agent frameworks) to securely discover and execute tools, read resources, and query live data from external platforms like Nodzo.

How do I get access to the Nodzo MCP Server?

The Nodzo MCP server is live in production at https://web.nodzo.ai/mcp. Company administrators can generate an MCP API key directly from the Nodzo web management dashboard under Settings → API & Integrations → Model Context Protocol.

Is customer conversation data safe when queried via MCP?

Yes. All MCP requests are authenticated using scoped API tokens over TLS 1.3 encryption. Nodzo enforces strict tenant data isolation, and customer voice data is never used to train public foundation models. Only authorized tools and resources accessible to your organizational role can be invoked.

Can I connect custom Python, LangChain, or CrewAI agents to Nodzo MCP?

Yes. The Nodzo MCP server supports standard Server-Sent Events (SSE) and HTTP streaming. You can connect custom agents using the official @modelcontextprotocol/sdk in TypeScript or Python, or any MCP-compatible client.

Does MCP support multilingual Hindi and English conversation transcripts?

Yes! All transcript tools return full dual-channel text with high-accuracy Indian accents, Hindi, English, and Hinglish code-switching, including speaker separation between rep and customer.

Available on All Plans & 14-Day Free Trial

Ready to Supercharge Your AI Workflows with Call Data?

Connect the live MCP endpoint today or book a personalized 20-minute onboarding demo to activate your 14-day free trial with 500 AI credits.

Open Live MCP Interface Book a Demo (14-Day Free Trial)