How Your Data Is Handled on getqueryly: Privacy, Purge, and What We Never Store

September 16, 2026

When you ask if getqueryly is safe to use, you are really asking three things: where does my file go, who can see it, and how long does it stay. The answer is simple by design. Your file is processed in memory and auto deleted when the session expires, never stored permanently. How long that is depends on your account, but every session expires from inactivity, not on a hidden warehouse.

The Short Answer

What Happens When You Upload

  1. You drop a CSV, Excel, PDF, JSON, or image. The file is validated, loaded into a temporary session, and shown as columns and row count in your workspace.
  2. You ask a question in plain English. The AI writes code, runs it in a isolated sandbox with strict time and memory limits, and returns charts and a written answer. The code runs apart from the main app, so a bad query cannot bring the service down.
  3. You see the result. The chart image and narrative live in that session. Share or export creates a snapshot you control, the original file is not copied to a warehouse.

Uploading is free. Asking a question costs 1 Make, health check 2, insights 3, multi analyst swarm 5 for quick or 8 for full. See live costs at GET /api/makes/costs.

What We Keep and What We Purge

WhatKept?For how long
Your raw fileTemporary session onlyAuto deleted with the session, after inactivity: anonymous one hour, free 24 hours, paid 7 days, sooner if you clear the session or delete the file. Every activity extends the window, so one hour means one hour of no activity, not one hour from upload.
Session results (charts, stats, answer)In sessionSame session lifetime, deleted with the session. A background job runs every five minutes to purge expired sessions, their messages, file records, and the actual files on disk.
Shared link snapshotOnly if you click ShareStored until you delete it, charts are embedded as images in that snapshot
Account and API keysYes, in your accountUntil you revoke or delete them
Usage count (Makes used)Yes, counters onlyFor billing and rate limits, not the file content

We do not keep a copy of your spreadsheet for later. If you need the same file tomorrow, you upload it again. That is intentional.

Who Can See Your Data

Only you in that session, and only while it is alive. We do not expose a public gallery of uploads and we do not send your file to a third party for storage. When AI writes code to answer your question, it sees column names and a sample of rows needed to build the query, not an open share. If you create a shared link with POST /api/data-scientist/share, that specific snapshot becomes viewable at /shared/{id} without sign in, because you chose to share it. Anything you do not share stays private to your account and expires with the session.

What To Use It For, and When To Pause

This is where the vibe code conversation matters. getqueryly shines when speed and plain English beat formal process:

That is not a flaw in the tool, it is a fit question. Use getqueryly where fast answers help, and keep regulated data on regulated systems.

Four Habits That Make Any AI Tool Safer

  1. Strip identifiers before upload. Replace names, emails, or IDs with codes if the question does not need them. You still get the chart without the risk.
  2. Use API keys with expiry and revoke. In Settings, generate a named key with 7 or 30 day expiry, rotate when a teammate leaves, revoke when a contract ends. Free 2 keys, starter 5, developer 10, business 25. See API keys done right.
  3. Track Makes and rate limits per key. Per key limits are 10 per minute on free, 30 starter, 60 developer, 120 business, half for AI endpoints. Check how to track Makes usage and avoid rate limits so a script does not burn quota or leak a key in logs.
  4. Share snapshots, not raw files. When you need to show a result, use Share to embed chart images in a view only link, instead of forwarding the original spreadsheet.

What We Fixed So You Do Not Have To Think About It

Recent hardening you benefit from without seeing it: raw errors are scrubbed so provider messages or file paths never reach the browser, failed code is retried automatically, watcher context prevents the AI from repeating a failed approach, and exports are built with fallbacks so a missing chart does not become a 500. You see a clean answer, the details stay in server logs.

If your data is a classroom survey or a shop sales sheet, upload and ask, you will have charts in seconds. If it is regulated data, sample and anonymize first. The tool respects the boundary you set.

Try It With Your Own File

Drop in any CSV or Excel, ask what is my best product by revenue or is this growth real or random noise, and compare the answer to what you would build by hand. For developers, the same flow is one header away: X-API-Key: YOUR_API_KEY for REST and MCP at https://getqueryly.com/mcp/sse, see how your AI assistant can actually do things.

Try getqueryly free, no credit card

Build note: getqueryly is built by a solo founder with AI assistance. That is why it ships fast and prices low. That is also why we publish this plain English handling guide and purge by default, so you know exactly what you are trusting before you upload.