fgaQ: 4 Tech questions
Does it work in Linux?
I understand that the Desktop utilizes OrcaSheet to generate the SQL query based on a prompt. After that, the app on the Desktop/Client connects to the database and retrieves the data without going through a third party. Is that correct?
Can I inspect, edit (before firing), and save the SQL query?
Database schemas commonly contain unmeaningful names, sometimes to shorten names or comply with naming rules. Can I transmit to OrcaSheet some kind of metadata, for example to map the physical column names to business names on a OLAP schema?
Navdeep_OrcaSheets
Jun 22, 2026A: Hello, thanks for the detailed questions. Here are our answers in order:
Linux: Yes. OrcaSheets is a desktop app and runs on Linux.
Query flow and third parties: You've got the part that matters most right. When a query runs, the app connects to your database directly and the data it returns stays on your machine. That data never passes through OrcaSheets or any third party. The one step that does reach out is the prompt-to-SQL translation: to turn your natural-language prompt into a query, OrcaSheets sends the prompt and your schema structure (table and column names, not the contents of any row) to a language model. So your data never leaves your environment, while the schema needed to write the query does.
Inspecting, editing and saving the SQL: Yes. You can review the generated SQL before it runs, edit it, and save it for reuse.
Mapping physical names to business names: We're enabling a knowledge base for your org that handles exactly this. You'll be able to map your physical column names to business-friendly names so the model uses them when writing queries against your OLAP schema. We'll have it enabled by tomorrow, and we'll let you know once it's live :)
Happy to go deeper on any of these.
Hi Navdeep. Thanks indeed for your reply. I will surely acquire the LTD this week for an OLAP application that we will be refactoring in August. It would also be great if I could import of the metadata (json or CSV format chosen by you) or access to a MCP tool that returns the mapping (business - column names) and the implicit relationships and referential columns.
Good news, the knowledge base is now live in the org admin portal. You can add your business terms, column definitions and table relationships directly, and the AI chat uses them to generate more accurate SQL in your own terminology. Here's how it works: https://orcasheets.ai/docs/knowledge-base. For your OLAP refactor in August, that means you can set the full mapping up as you go.