Skip to main content
Glama
Tunzaa MCP Server

Tunzaa MCP Server

Official
by Tunzaa

Malipo MCP Server

Grounding for AI-Driven Payment Integrations

The Malipo MCP Server is a developer-centric tool built for the community. It provides high-fidelity grounding data, integrated documentation, and "Golden" code patterns that allow AI agents (vibe coders) to generate perfect, non-hallucinated integration code for the Malipo by Tunzaa ecosystem.

🚀 Instant Start (Fastest Way)

You can run the server directly from GitHub without cloning or installing dependencies.

1. Claude Desktop

Add this to your claude_desktop_config.json:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

  • Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "malipo": {
      "command": "npx",
      "args": ["-y", "github:Tunzaa/tunzaa_mcp"]
    }
  }
}

2. Cursor

  1. Go to Settings -> Features -> MCP.

  2. Click + Add New MCP Server.

  3. Name: Malipo | Type: command | Value: npx -y github:Tunzaa/tunzaa_mcp

3. Windsurf

Add this to your ~/.codeium/config.json:

{
  "mcpServers": {
    "malipo": {
      "command": "npx",
      "args": ["-y", "github:Tunzaa/tunzaa_mcp"]
    }
  }
}

Related MCP server: Yolfi MCP Server

🧭 The Vibe Coding Workflow

This server is designed to help you build Malipo integrations in minutes. Follow this flow with your AI assistant:

  1. Grounding: Add this MCP server to your project.

  2. Exploration: Ask the AI: "List the Malipo resources and read the authentication guide."

  3. Simulation: Run the tool: create_demo_shop to see a live trace of a successful integration.

  4. Generation: Ask the AI: "Based on the grounding trace and the node-express example, build a checkout page for my app."


🏪 The Grounding "Demo Shop"

The create_demo_shop tool is the cornerstone of this platform. It doesn't just return data; it provides a Live Grounding Trace.

How to use it:

  1. Trigger the Simulation: Tell your AI agent: "Run the Malipo create_demo_shop tool to understand the payment flow."

  2. Review the Trace: The agent will receive a chronological sequence of calls including Authentication, Payment Initiation, and Installment creation.

  3. Production Implementation: Each step in the trace contains "Grounding Insights" that teach the agent how to handle state, headers, and reference IDs in your actual code.

  4. Boilerplate: Ask the agent to "Convert the grounding trace into a [Node/Python/PHP] implementation using the best practices found in the documentation resources."


✨ Features

  • Integrated Documentation: AI agents can "read" guides on Auth, Payments, and Webhooks directly through MCP Resources.

  • Golden Patterns: Embedded code snippets for Express.js, React Hooks, and more.

  • Vibe Coder Optimized: Rich schema descriptions and instructional traces (via create_demo_shop) ensure zero hallucination.

  • Mock Mode by Default: Generates "Golden" mock data matching the real Malipo API structure.

  • Live Mode (Optional): Real-time verification against the Malipo Sandbox/Production.


🛠️ Usage (Live Mode)

To have the AI verify real data from your Malipo account (e.g., checking transaction statuses), add your credentials to the env block in your config:

"env": {
  "MALIPO_API_KEY": "your_api_key",
  "MALIPO_SECRET_KEY": "your_secret_key",
  "MALIPO_ENVIRONMENT": "sandbox"
}

Legacy TUNZAA_* environment variables are still accepted as a fallback.

Optional settings

Variable

Default

Purpose

MALIPO_API_BASE_URL

https://pay.tunzaa.co.tz

API host. Sandbox and production share this host and are selected by MALIPO_ENVIRONMENT.

MALIPO_ALLOWED_HOSTS

(empty)

Comma-separated extra hosts that the tools' address / api_url override may target, e.g. staging.example.com,localhost:8000. Overrides must use HTTPS (plain HTTP only for localhost). Any other host is rejected before a request is made.

MALIPO_EXPOSE_TOKEN

false

get_token returns a masked token (eyJh…x9Q) and its expiry. Set to true to return the full bearer token.

create_demo_shop initiates a real payment and creates a real installment plan, so it only runs when MALIPO_ENVIRONMENT=sandbox (or in mock mode).


🏗️ Local Development

If you'd like to contribute or modify the server:

  1. git clone https://github.com/Tunzaa/tunzaa_mcp.git

  2. cd tunzaa_mcp && pnpm install && pnpm run build

  3. Use the local path in your config: "args": ["/ABSOLUTE/PATH/TO/tunzaa_mcp/dist/index.js"]

License

ISC

Available Tools

10 tools
create_demo_shopC

The ultimate grounding tool. Executes a full sequence of Malipo API calls (Token -> Payment -> Installments). Use this to see a 'live trace' of the API, allowing you to generate perfect integration code.

Inputs: api_url.

create_installmentB

Create a new installment plan. Use this to understand the complex object structure required for installment-based payments.

Inputs: name (required), customer (required), end_date (required), start_date (required), +5 more.

delete_installment_planC

Cancel/Delete an existing installment plan. Use this to verify the cancellation response structure.

Inputs: plan_id (required), address.

edit_installment_planC

Update an existing installment plan. Use this to understand which fields are mutable via the Malipo API.

Inputs: plan_id (required), updates (required), address.

get_installment_planB

Get details of a specific installment plan. Use this to see the precise fields returned for a plan (status, schedules, etc.).

Inputs: plan_id (required), address, include_payments.

get_payment_statusB

Check the status of a payment transaction. Helpful for understanding the various status states (COMPLETED, PENDING, FAILED) for your application logic.

Inputs: transactionID (required), address.

get_tokenB

Retrieve a Malipo API access token. Refreshes internal token automatically. Use this to verify your API credentials and see the internal token structure.

Inputs: address.

handle_callbackC

Simulate or handle the callback payload sent by Malipo to your webhook. Essential for grounding your webhook integration code with real payload examples.

Inputs: status (required), transaction_id (required), amount, timestamp, +3 more.

initiate_paymentC

Initiate a payment request (M-Pesa, etc.) via Malipo API. Call this to inspect the response structure needed to implement mobile money flows in your local code.

Inputs: amount (required), reference (required), customer_msisdn (required), address, +1 more.

list_installmentsB

List existing installment plans. Use this to see how pagination and plan summaries are returned by the Malipo API.

Inputs: from, limit, address.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3.3/5.0

    Scored across 10 tools

    Disambiguation4/5

    Most tools target distinct actions and resources: auth, payment initiation, callback handling, installment CRUD, and payment status. create_demo_shop is a composite trace tool that overlaps with get_token, initiate_payment, and create_installment, and its name does not perfectly match its grounding purpose, but it remains distinguishable.

    Naming Consistency5/5

    All names use consistent snake_case verb_noun patterns, such as get_token, initiate_payment, list_installments, and create_installment. Verbs like get, list, create, edit, and delete are used predictably across the set.

    Tool Count5/5

    The set has 10 tools, which fits the payment and installment API domain well. Each tool appears to earn its place by covering a distinct operation or grounding scenario.

    Completeness4/5

    Installment plan coverage includes full CRUD plus listing, and payment coverage includes initiation, status, callback handling, and token retrieval. Minor gaps exist on the payment side, such as listing payments, retrieving full payment details, or refund/cancel operations, but agents can largely work around these.

    Maintenance

    ActivitySlowing
    ResponsivenessLimited data

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      Not graded
      quality
      C
      maintenance
      Enables AI coding agents to safely integrate cross-border payments by providing structured payment knowledge, deterministic validation of idempotency, security, and state-machine rules, webhook testing, executable pytest generation, and diagnosis for Python/FastAPI Checkout projects.
      -
    • F
      license
      Not graded
      quality
      A
      maintenance
      Enables AI coding agents to integrate stablecoin checkout, payment links, payment status checks, webhook verification, and webhook-based entitlement logic into applications.
      -