Sign in Subscribe

Your Spreadsheet Is Already an App. DataBased.si Just Makes It AI-Native.

Your business runs on spreadsheets — whether you admit it or not

Customer lists. Sales pipelines. Inventory. Project trackers. Financial models. For all the hype around CRMs, data warehouses, and "digital transformation," the operational heart of most teams still beats in an Excel file or a Google Sheet somewhere.

And honestly? That's fine. Spreadsheets are flexible, familiar, and everyone already knows how to use them. The problem isn't the spreadsheet — it's that spreadsheets were never designed to talk to AI.

That's the gap DataBased.si set out to close.

What is DataBased.si?

DataBased is a bridge between the data you already have and the AI tools you already use. Connect an Excel workbook, a Google Sheet, or a CSV/XLS file, and DataBased turns it into a structured, AI-accessible data source through the Model Context Protocol (MCP).

In plain terms: instead of exporting, copying, or attaching spreadsheets to a chat window and hoping the model figures out what column B means, DataBased.si gives AI models a real, structured understanding of your data — what the columns represent, how entities relate, which values are currencies, dates, or pipeline stages.

Then you can just ask questions. And AI can answer — or even write back.

From spreadsheet to AI app in three steps

The workflow is deliberately simple:

1. Connect your data. Link a Google Sheet or Excel file, or upload a CSV/XLS. No migration, no new database, no replatforming.

2. DataBased.si understands it. The service analyzes your columns, headers, and data to build a semantic model — recognizing that "Revenue" is a currency field, "Status" is an enum with four pipeline stages, and "Customer" is the primary key of a customers entity.

3. Give AI access. Your data goes live as an MCP server. Any compatible AI client — ChatGPT, Claude, Cursor, custom agents — can discover it, query it, and interact with it.

That last part is the key. Because DataBased.si exposes your data through MCP, you're not locked into one AI tool. One dataset, every AI interface.

Ask your spreadsheet anything

Once connected, working with your data looks like this:

  • "Which customers haven't been contacted in over 30 days?"
  • "What's our total pipeline value for deals in the negotiation stage?"
  • "Which SKUs are below their reorder threshold?"

No formulas, no filters, no scrolling through hundreds of rows. Just plain-language questions and answers grounded in your actual data.

It doesn't just read — it writes back

Here's where DataBased.si goes beyond "chat with your spreadsheet" gimmicks. With write-back enabled, AI can update your spreadsheet directly:

  • Log a follow-up action after a customer analysis
  • Update a deal's status in your pipeline
  • Add new records to an inventory list

Your spreadsheet remains the single source of truth — AI just becomes another (very fast) pair of hands working in it. And because DataBased.si separates read and write permissions, you decide exactly what AI is allowed to touch.

Built on an open standard

For developers, this is the most interesting part. DataBased.si is built around MCP, the emerging open protocol for connecting AI models to external tools and data. Each connected dataset gets its own MCP server with discoverable tools and resources — think search_customers, get_customer, update_customer, add_record.

That means any MCP-compatible client can integrate your data natively, without custom APIs or glue code. It's a clean, standard interface between your data and whatever AI stack you run.

Your data, your control

Handing AI access to business data raises obvious questions, and DataBased.si addresses them head-on:

  • Controlled access — you choose which spreadsheets, sheets, and columns are exposed
  • Permission-aware workflows — read and write access are configured separately
  • Transparent interactions — every read and write performed through MCP is visible
  • Built for real operations — designed for data teams depend on daily, not throwaway demos

Who is it for?

Practically anyone whose work lives in a spreadsheet: sales teams querying pipelines, operations staff updating records, finance teams exploring budgets, customer success managers adding follow-ups, researchers analyzing datasets, or developers turning internal spreadsheets into lightweight AI applications — without building anything.

Wrapping up

The pitch is refreshingly non-disruptive: keep your spreadsheet, keep your workflow, keep your data where it is — and add an AI layer on top. No migration. No retraining your team. No rebuilding tools that already work.

If you've been waiting for a way to bring AI into your existing data workflows without a six-month migration project, DataBased.si is worth a look. There's a free tier to get started, and no credit card is required.

Try it at databased.si

Pricing note: DataBased.si lists Free, Pro, and Business plans, with final pricing to be announced at launch.

I reviewed the product's website and wrote a blog article, now open in the canvas.

What the article covers:

  • The hook — businesses still run on spreadsheets, but spreadsheets weren't designed for AI
  • The product — DataBased.si connects Excel, Google Sheets, and CSV/XLS files, then exposes them to AI models via MCP
  • The workflow — connect → DataBased.si builds a semantic model of your data → AI clients can query and even write back
  • Key differentiators — read/write-back, permission controls, transparent logging, and the open MCP standard (works with ChatGPT, Claude, Cursor, custom agents)
  • A closing pitch with the free tier and a pricing caveat, since final pricing isn't announced yet

All facts come from the product's own site, so I'd recommend a fact-check with the team before publishing (especially pricing and MCP client availability). Want me to adjust the tone (more technical / more casual), shorten it, or add SEO metadata and a meta description?