Skip to main content
Field notes / philosophy

What you keep if you leave

Plexara queries your data where it lives and never moves it into a store of its own. What people make on it are ordinary files, and what your team writes is readable text. Portability is a property of where things are kept, not a promise on a slide.

7-minute readPhilosophy

What accumulates on a data platform

The data was always yours. It sits in your warehouse, your object storage, your operational databases. What a data platform adds on top is meaning and work: the description that says what a column actually holds, the glossary term that fixes a business definition, the report a team reads every Monday, the file an automation refreshes every night.

That layer accumulates. On Plexara it accumulates as a byproduct of use, because documentation happens while people work rather than in a separate cataloging project. Over months it becomes the most valuable part of a deployment.

So the lock-in question has two halves. Where does the data have to live for the platform to use it? And what form does the work take once it is done?

What accumulates

Descriptions

What a table or column actually means

Glossary terms

The business vocabulary your team agreed on

Lineage

Which work produced which table or file

Tags

Sensitivity, ownership area, status

Ownership

Who is accountable for each dataset

Reports and files

What people and automations produce

Every conversation that explains a dataset, and every report someone saves, leaves something behind. This is the layer that grows with use.

The data stays in your systems

Plexara reaches your data through connections. A warehouse table is queried in the warehouse, a CRM record is read through the CRM’s API, and a question that joins the two federates across both at the moment it is asked. Nothing is copied into a store of Plexara’s own to make it queryable, so there is nothing to extract when you stop.

A spreadsheet someone uploads becomes a table over the file itself, and the file remains a CSV. A result your agent or an automation exports is written as CSV, JSON lines or Parquet, open formats that every query engine and spreadsheet reads, with the column types intact in the Parquet case.

This is a property of the architecture rather than a policy. A platform that must ingest your data before it can answer questions owns a copy that has to be retrieved. A platform that reads data where it lives only ever held a connection.

Where the data has to live

Ingest first, then answer

Copied into the platform

  • Data loaded into a store the vendor runs before it can be queried
  • A second copy to keep in step with the source
  • Leaving means retrieving that copy in whatever form the vendor offers
  • The switching cost points back at the vendor

Connections, not copies

Read where it lives

  • Warehouse tables queried in the warehouse, APIs read through the API
  • Uploaded spreadsheets stay files, registered as tables over the file
  • Results exported as CSV, JSON lines or Parquet
  • Your sources keep their data; the platform held a connection

The work is ordinary files and readable text

The reports, dashboards and decks people make on Plexara are HTML, SVG and Markdown files. Each one has a Download button beside Share, and a report or deck exports to PDF as it looks on screen. There is no proprietary container between your team and what it built.

The meaning your team writes down is plain text: descriptions and glossary definitions attached to the tables they describe, context notes and knowledge pages in Markdown, readable by anyone with access in the portal. It is written in your vocabulary, about your data, by your people.

It is the same reasoning behind building on protocols rather than products. Agents reach Plexara over an open protocol, queries run through an open engine, and results land in open file formats, so the layer in the middle can evolve or be replaced without holding the rest hostage.

What stays yours

In your systems

  • Warehouse and database tables
  • Files in your storage
  • Records in your SaaS tools

Ordinary files

  • Reports and dashboards as HTML
  • Decks and documents, with PDF export
  • CSV, JSON lines and Parquet exports

Readable text

  • Descriptions and glossary definitions
  • Context notes on tables
  • Knowledge pages in Markdown
Nothing in the left column ever moved. The middle column downloads as the files it already is. The right column is your team’s own writing, readable in the portal by anyone with access.
The Assets page in the portal: Assets and Collections tabs, Mine, Shared, and All scopes, a search box, type and tag filters, a sort control, a grid or list toggle, and a grid of asset cards including a JSON-lines file, a Q3 regional review slide deck, an ACME sales dashboard, a data quality summary, and a Q4 revenue dashboard, each with a preview, a description, a content type badge, tags, a size, and a date.
Every saved asset is an ordinary file, and each card says which kind: HTML, SVG, Markdown, JSON lines. There is no proprietary container between you and the report.

Portability is a property, not a promise

Keeping what you built should not require staying. When the data never moved in, and the work is files in formats any tool reads, continuing with a platform becomes a decision about the value it delivers, not one forced by the cost of leaving.

Regulators have encoded the same instinct for personal data: the right to receive it in a structured, machine-readable format and move it elsewhere without hindrance. It is a sound standard to hold any data platform to. Ask where your data has to live for it to work, and what form the work takes when it is done.

The asset viewer open on a Q4 revenue dashboard with Download and Share buttons in the header and the Share dialog in front: people with access and their roles, the Data Analyst persona as a group, general access set to anyone who can sign in, a post to a Mattermost channel, a pending access request, and Copy link and Post to a channel buttons.
Download sits next to Share. The dashboard leaves as the file it is, to a teammate through a share or to a laptop through the Download button, and either way it is the same file.