Didit MCP

Product Introduction

1. Definition: Didit MCP is an official Model Context Protocol (MCP) server that connects AI coding assistants and chat clients—such as Claude, ChatGPT, Cursor, and VS Code—directly to the Didit identity, compliance, and fraud prevention platform. Operating as a hosted HTTP server (transport: Streamable HTTP) located at https://mcp.didit.me/mcp, it exposes 115 first-class identity and fraud tools spanning 19 technical domains without requiring manual API key configuration.

2. Core Value Proposition: The product exists to bridge natural language AI agents with enterprise-grade identity verification, transforming complex API integrations into conversational commands. By leveraging OAuth 2.1 with PKCE for secure authentication, Didit MCP allows developers, compliance officers, and system administrators to build KYC (Know Your Customer) workflows, run KYB (Know Your Business) registry lookups, screen wallets and individuals against AML (Anti-Money Laundering) databases, and triage fraud review queues entirely through natural language prompts.

Main Features

1. Natural Language Workflow Orchestration: The MCP server allows users to construct complex verification pipelines by simply instructing their AI agent. Using tools like didit_workflow_create, the agent dynamically composes multi-step verification flows—combining ID document validation, biometric liveness checks, and government database validation—complete with conditional branching and automated decisioning matrices (approved, declined, in review).

2. Comprehensive AML and Wallet Screening: Exposing real-time compliance checks via tools like didit_aml_screening and didit_transaction_screen_wallet, the server evaluates individuals, corporate entities, and cryptocurrency wallets against global sanction lists (OFAC SDN, EU Consolidated, UN Security Council), Politically Exposed Person (PEP) tiers, adverse media, and high-risk blockchain exposure vectors, returning structured, risk-scored JSON outputs.

3. Case Management and Review Queue Triage: The platform integrates manual compliance reviews directly into the AI agent workspace. Tools such as didit_review_queue_list and didit_case_create enable agents to scan flagged sessions (e.g., triggered by document glare or low face-match confidence scores), open investigation cases, append audit notes, and assign review items to compliance teams based on granular console roles.

Problems Solved

1. Pain Point: Integrating identity verification and anti-money laundering compliance traditionally requires writing extensive boilerplate API integration code, managing secure webhook endpoints, and context-switching between developer environments and third-party dashboards. Didit MCP eliminates these development friction points by turning 115 platform capabilities into native agentic tools accessible via natural language.

2. Target Audience: Full-stack software engineers, fintech compliance officers, risk operations managers, blockchain developers, and technical founders implementing KYC/KYB infrastructure in applications spanning 220+ countries and territories.

3. Use Cases: Rapidly scaffolding onboarding flows in IDEs like Cursor or VS Code; executing on-demand wallet sanctions screening prior to cryptocurrency withdrawals; bulk triaging compliance review queues using conversational AI in Claude Desktop; and querying credit balances and monthly conversion analytics without navigating away from the development workspace.

Unique Advantages

1. Strengths & Limitations (Pros & Cons):

   • Pros: Zero API keys or secrets required in local config files due to OAuth 2.1 PKCE authentication; native support across diverse MCP clients (Claude Code, Claude Desktop, ChatGPT Developer Mode, Cursor, VS Code); $10 of free monthly credit with identical per-success pricing to the core API; open-source core available on GitHub for self-hosting options.
   • Cons: Relies on external hosted infrastructure (mcp.didit.me/mcp) for turnkey operation; requires active internet connectivity for OAuth authentication; advanced multi-tenant enterprise configurations may require manual console interventions beyond conversational scope.

2. Key Alternatives & Differentiation: Unlike traditional Identity-as-a-Service (IDaaS) platforms that force developers to manually read API documentation and write custom wrapper functions, Didit MCP natively embeds compliance infrastructure directly into the developer's AI workflow. Compared to building custom internal scripts around REST APIs, Didit MCP provides an out-of-the-box, standardized protocol layer with 115 pre-built tools, role-based access control, and built-in regulatory alignment across 220+ global jurisdictions.

Frequently Asked Questions (FAQ)

1. How do I connect Didit MCP to Claude Code or Cursor? For Claude Code, execute the command claude mcp add --transport http didit https://mcp.didit.me/mcp followed by /mcp to complete the browser-based OAuth login. For Cursor, add the server configuration object {"mcpServers": {"didit": {"url": "https://mcp.didit.me/mcp"}}} to your ~/.cursor/mcp.json file.

2. Do I need to store API keys in my configuration files? No. The Didit MCP server utilizes OAuth 2.1 with PKCE for secure authentication through the Business Console. The AI agent acts directly as the signed-in user, respecting their assigned role permissions (Owner, Admin, Compliance Officer, Developer, Reader) without exposing API keys or secrets in local config files.

3. How is Didit MCP priced? The MCP layer itself is completely free. Users receive $10 of free credit every month with no credit card required. Beyond the free tier, verifications and screenings are billed at standard public per-success API prices (starting from $0.33 per full KYC check and $0.15 per wallet screen).