Marqeta DiVA

You run an MCP server that exposes Marqeta DiVA data through a programmable interface. This server lets you fetch transaction, balance, card, user, and chargeback data, perform flexible queries, export results, and optionally enable local storage and semantic features for advanced analytics. You can connect it to Claude Desktop or other MCP clients to get reporting, dashboards, and AI-assisted insights from your DiVA data.

How to use

Start by running the MCP server locally using your preferred method. You will then connect an MCP client to the running server and issue data-queries that match your reporting needs. You can fetch authorizations, settlements, clearings, declines, and loads; retrieve program and activity balances; obtain card and user details; access chargeback information; and discover available views and schemas. You can filter, sort, select fields, and specify date ranges to tailor results. If you enable RAG features, you gain local storage, semantic search, and similarity-based lookups, which can help when API rate limits or token constraints apply.

How to install

Prerequisites: you need Python 3.10 or higher and the uvx runtime for convenient local execution. You also need Marqeta DiVA API credentials: Application Token, Access Token, and Program Name.

Option 1: Run with uvx (recommended). No installation is required beyond having uvx installed. Basic features only: run the server with the following command.

uvx marqeta-diva-mcp

Option 1.1: Run with uvx and enable RAG features (local storage plus semantic search). This loads the extended capabilities.

uvx --with marqeta-diva-mcp[rag] marqeta-diva-mcp

Option 2: Traditional installation. Core features only.

pip install marqeta-diva-mcp

Option 2.1: With RAG features (recommended for advanced analytics).

pip install marqeta-diva-mcp[rag]

Option 3: From source.

cd marqeta-diva-mcp
pip install -e .              # Basic features
pip install -e ".[rag]"       # With RAG features

Additional configuration and usage notes

Create an environment file and supply your credentials to allow the server to access the DiVA API.

# Example environment file content
MARQETA_APP_TOKEN=your_application_token_here
MARQETA_ACCESS_TOKEN=your_access_token_here
MARQETA_PROGRAM=your_program_name_here

Starting and connecting to Claude Desktop or other MCP clients

Use the included MCP configuration examples to connect your client to the running server. The configurations show how to run the server and how to provide credentials.

If you enable local storage or RAG features, you will see a log message indicating that these features are ENABLED. If not, you will see a message indicating that they are DISABLED.

Security and access considerations

Protect your credentials. Do not expose your application token, access token, or program name in public or shared configurations. Use environment variables and proper access controls when connecting MCP clients.

Available tools

get_authorizations

Retrieve detailed authorization transaction data with amounts, counts, involved users/cards, and merchant information, with options to filter, sort, and aggregate.

get_settlements

Fetch settlement-related transaction data including status, post dates, amounts, and network information with filtering and sorting.

get_clearings

Obtain accounting-level line items for the transaction lifecycle to aid reconciliation.

get_declines

Retrieve declined transaction data including tokens, reasons, merchant info, and amounts.

get_loads

Access load transaction data with amounts and transactional details.

get_program_balances

Get program-level balances including starting/ending balances and transfer amounts.

get_program_balances_settlement

Get settlement-based program balance data and fund transfers.

get_activity_balances

Fetch cardholder-level balance data, expandable by network.

get_cards

Get card detail data including user tokens, card state, and UAI.

get_users

Get user detail data including tokens, UAI, and card counts.

get_chargebacks_status

Get chargeback status data including state and provisional credits.

get_chargebacks_detail

Get detailed chargeback information with dates and types.

get_transaction_token

Map Core API transaction tokens to DiVA report tokens for reconciliation.

get_card_counts

Get aggregated card counts over time (daily/weekly/monthly).

get_user_counts

Get aggregated user counts over time (daily/weekly/monthly).

get_activity_balances_network_detail

Get activity balances broken out by card network for network-level insights.

list_available_views

List all available DiVA API views with their metadata.

get_view_schema

Return the schema for a specific DiVA view including field names and descriptions.