How to Manage Your App's Data with Claude and nuzur's MCP Server
There's a problem that almost every small team runs into eventually: the developer sets up the data model, connects it to a database, and then becomes the bottleneck for every single data change.
Need to add a new record to a content platform? Ask the developer. Need to update a record that has the wrong value? Ask the developer. Need to add a new location entry? You guessed it — ask the developer.
This isn't a people problem. It's a workflow problem. And it's exactly what nuzur was built to solve.
Check out this short video I did on this:
The gap between your data model and your team
Most tools treat data management as a developer-only concern. You either write SQL directly, use an admin panel bolted onto your framework, or export to a spreadsheet and hope nobody breaks the formatting.
None of these work well for teams where not everyone is technical. And none of them give you confidence that a change is correct before it hits your production database.
nuzur takes a different approach: model your data visually, connect it to your database, and let your whole team interact with it through a structured interface — with a review layer before anything gets committed.
Recently, I took that a step further by introducing an MCP (Model Context Protocol) server. And it offers yet another way to interact with your data safely.
What is MCP and why does it matter?
MCP is an open standard that allows AI assistants like Claude to connect to external tools and data sources. Instead of copy-pasting your database schema into a chat window and hoping Claude understands it, an MCP server gives Claude a live, structured connection to your actual system.
With nuzur's MCP server, Claude can:
- Read your full project structure — every entity, field and relationship
- Always get the latest published changes as soon as they become available
- Understand how your data is connected across your model
- Create change requests on your behalf, pre-filled with the right values
That last point is important. Claude doesn't write directly to your database. It creates a change request — a proposed record that sits in review until a human approves it. This means you get the speed and intelligence of AI-assisted data entry, without giving up control.
A real example: Squirrels and Friends
To make this concrete, here's how this works in practice using Squirrels and Friends, a drag show tracking platform built on nuzur.
The project has 20+ entities and 50+ relationships — drag performers, shows, seasons, episodes, fantasy seasons, locations, and more. It's a real production app with real data.
Without nuzur's MCP integration, adding new records takes time, although the nuzur UI has all the required validations to ensure the data quality, it is time consuming to look up the data and format it accordingly, with the MCP server it takes a fraction of the time as Claude is able to fetch the necessary information and format it.
With the MCP server connected to Claude, the conversation looks like this:
You: Add a new episode — Season 18, Episode 3, aired last Friday, here is the raw html data.
Claude: I've created a change request for Episode 3 of Season 18. Here are the details I've filled in based on your model: name, number, air date, season reference, and status. The request is ready for review in nuzur.
Claude mapped natural language to the exact fields in the data model, correctly associated the episode with the right season, and formatted the date. No schema lookup required. No risk of a typo landing directly in the database. To achieve this Claude is able to query the data in the database, to find references to other entities or look at the format of existing records helping keep the data consistent.
Back in nuzur, the change request appears in the review queue. Anyone with the right permissions can open it, verify the details, and approve it. Only then does the record get written.
Why the review step matters
It might be tempting to skip the review and let Claude write directly to the database. But the review step is a feature, not a limitation.
AI models are powerful, but they make assumptions. A human reviewer catches the moment Claude misunderstood "last Friday" because the conversation happened on a Monday. Or when a field value looks right but conflicts with an existing record.
The change request workflow also creates a natural audit trail. You can see who requested a change, when, and who approved it. For teams managing shared data — content, products, customer records — that history is genuinely valuable.
Who this workflow is for
This isn't just for developers. In fact, the whole point is to make data management accessible to everyone on a team.
- Developers design and version the data model in nuzur's visual editor
- Data managers, operations staff, or analysts use Claude to propose data changes in plain language
- Team leads or reviewers approve or reject changes before they go live
The developer becomes an architect rather than a data entry bottleneck. The rest of the team gains autonomy without gaining the ability to accidentally break something.
Getting started
nuzur is free for personal use. If you want to try the MCP integration with Claude, you can connect the nuzur MCP server from your Claude settings and point it to the nuzur mcp server: https://ccmcp.nuzur.com you will need to authenticate with your nuzur account and you will be all set.
Your data model becomes something Claude can genuinely understand — not just read, but reason about. And your team gets a workflow that's fast, safe, and actually built for collaboration.