Data & Analytics · MotherDuck MCP + Claude
MotherDuck MCP with Claude: Query and Explore Your Data in Plain Language
The short answer
Claude works with MotherDuck through MCP, the Model Context Protocol. MotherDuck's official hosted MCP server acts as a secure bridge. Once connected, Claude can explore schemas, query data, manage Dives and Guides, and run SQL writes from natural language requests, within permissions your organisation sets.
What is MotherDuck MCP?
MotherDuck is a cloud data platform built around DuckDB, a fast analytical database. Teams use it to store, query and share data for reporting and analysis. Analysts, data engineers and business teams use it to answer questions quickly. MotherDuck MCP is the company's official hosted server for database access and management, giving AI tools a governed route into that data.
How Claude connects to MotherDuck MCP through MCP
The Model Context Protocol is an open standard that lets Claude securely use other software. Think of it as a universal adapter between Claude and a tool. MotherDuck provides its own official hosted MCP server, so Claude connects to MotherDuck directly rather than through a workaround.
Once connected, Claude can look at the schemas and tables available to it, query data using natural language, and help manage Dives and Guides. It can also run SQL writes when the task calls for them. Claude turns a plain question such as which regions grew fastest last quarter into the right query and explains the result.
Permissions and approvals keep people in charge. Sign-in uses Dynamic Client Registration OAuth, so access is granted through the account owner's existing MotherDuck sign-in rather than shared passwords. Access can be limited to the databases and actions each person needs, and write actions can require a person's approval before they run.
What you can automate
Self-service reporting for managers
A regional sales manager asks, in plain English, for the week's pipeline by product line. Claude explores the relevant schema, runs the query and returns a summary table. The manager gets answers in minutes instead of waiting for a ticket to be processed by the data team.
Data quality checks
A data engineer asks Claude to scan a set of tables for missing values, duplicate keys or unexpected date ranges. Claude runs the checks and lists the issues it finds, with the queries used, so the engineer can review and fix problems before they reach dashboards.
Maintaining Dives and Guides
A analytics lead asks Claude to update an existing Dive so it reflects a new metric definition, or to draft a Guide that explains how a report should be read. Claude makes the changes in MotherDuck, and the lead reviews them before the material is shared with the wider team.
Finance month-end support
A finance controller asks Claude to reconcile transaction totals across two tables and list any differences above an agreed threshold. Claude prepares the queries and the difference report. Finance staff check the results and decide on any adjustments.
Controlled data updates
An operations lead needs a batch correction to a set of records. Claude drafts the SQL write, shows the statement and the rows it would affect, and waits for a named approver to confirm before the change runs. This keeps routine fixes fast while protecting the data.
An end-to-end workflow: Weekly customer health report
TriggerEvery Monday morning, or when a customer success manager asks for an update
Claude queries MotherDuck for product usage, support ticket counts and renewal dates for the past seven days
Claude pulls account owner details and open opportunities from the CRM through a connected system
Claude flags accounts with falling usage, rising tickets or renewals within 60 days
The results are written into a Guide in MotherDuck with a plain-language summary for each flagged account
The customer success manager reviews the flagged list and assigns follow-up tasks in the CRM
OutcomeCustomer success teams start each week with a clear, reviewed list of at-risk accounts, typically saving hours of manual compilation and catching churn signals earlier.
How we set it up
Free automation audit
We review your current reporting and data requests with you and identify the workflows where Claude and MotherDuck would save the most time.
Access design
We set up sign-in using MotherDuck's Dynamic Client Registration OAuth, so each user connects with their own account and sees only the databases they are allowed to use.
Workflow build
We design the end-to-end workflow across MotherDuck and your other systems, such as Microsoft 365, CRM or ticketing, and agree which actions need human approval.
Testing with real cases
We test the workflow against past questions, reports and data corrections, and check that results match what your analysts expect before anyone relies on them.
Launch and training
We launch with monitoring and audit logs turned on, then train your team on how to ask good questions, review results and handle approvals.
Who benefits
- Data and analytics leads
- Less time spent on routine query requests, with Dives and Guides kept current through reviewed changes.
- Business managers
- Faster answers to operational questions without waiting for a data team queue.
- Finance and operations teams
- Reconciliations and checks prepared quickly, with a clear record of what was run and who approved it.
- IT and security heads
- Sign-in through existing accounts, limited access and audit logs that support oversight of data use.
Security and control
- Sign-in through Dynamic Client Registration OAuth, so no shared passwords are stored in the AI setup
- Least-privilege access, limiting each user to the databases and actions their role requires
- Human approval for SQL writes and other consequential changes before they run
- Audit logs and monitoring to record what was queried, changed and by whom
- Staged testing against real cases before launch, with a rollback plan for any write process
For teams in the US, UK, India and the Middle East
Organisations in the United States and the United Kingdom can host their Claude-based workflows in those regions to align with local data protection expectations. Indian organisations are supported under the Digital Personal Data Protection Act, 2023, and we design access and retention to match its requirements. Organisations in the UAE and Saudi Arabia are served with attention to local data protection laws. Hosting region is chosen during design to meet your specific legal and contractual requirements.
Learn more about Claude and MCP
Anthropic's AI model and assistant.
Secure links between Claude and business tools.
How engineers build and change software with Claude.
Hosted agents that run multi-step work.
The open standard Claude uses to connect to tools.
Frequently asked questions
What does the MotherDuck MCP server do?
MotherDuck MCP is the company's official hosted server. It gives AI tools a secure route into MotherDuck data, schemas and database operations. It lets an approved assistant explore what data is available, run queries, manage Dives and Guides, and carry out SQL writes when permitted. It does not replace your data governance; it works within the permissions you set.
Is it secure to connect Claude to MotherDuck?
Security depends on how the connection is set up. MotherDuck MCP uses Dynamic Client Registration OAuth, so users sign in with their own accounts instead of sharing credentials. We typically add least-privilege access, approval steps for write actions and audit logs. We test these controls before launch, though no setup can guarantee complete security.
What can Claude actually do with MotherDuck?
Once connected, Claude can explore schemas, query data from plain-language requests, help manage Dives and Guides, and run SQL writes. Typical uses include reporting, data checks, reconciliations and controlled corrections. The exact actions available depend on the permissions granted to each user.
How long does setup take?
Setup time depends on the number of workflows and systems involved. A focused first workflow is often ready within a few weeks, including the free audit, access design, testing and training. Larger programmes covering several teams and systems take longer. We agree a timeline after the audit.
What plan or permissions are needed?
You need an active MotherDuck account with the databases your team uses, and an account owner or administrator who can approve access. The specific MotherDuck plan that suits your usage is a question for your MotherDuck account contact. We confirm the permissions your workflow needs during the audit.