Every security answer, on one page.
Whoever signs off your vendors has a list of questions, and they are answered below rather than in a call: what we encrypt and what the disk encrypts, what we store and what we do not, exactly what leaves for the model, and what happens on the SQL. Intellrise runs on your own AI provider key — so the model call is billed to your account, under your contract — attaches PostgreSQL, MySQL and Redshift sources read-only, and shows you the exact SQL that ran. On an ordinary question what reaches the model is your question, your table and column names, the SQL, and the first ten rows of each result so the written answer matches the numbers; the rest of the result is not sent, and the one other case that sends rows is named further down. What we keep is the answer, in your account, deletable.
What we encrypt, and what the disk encrypts
Two layers, and they are not the same promise, so a vendor review should have both. We encrypt your data-source credentials and your AI key ourselves, with AES-256-GCM, before they reach the database — your key is only ever shown back to you masked. Everything else we keep — your chat history including the rows each answer returned, the reports and dashboards you save, and the files you upload — is stored in our database, encrypted at rest with AES-256 and reached only over TLS. Where it physically sits: Singapore.
Per-account isolation
One shared database in which every row carries the account it belongs to and every query filters on it. Section 6 of our Privacy Policy describes the same mechanism.
We keep the answer, not your database
We keep no standing copy of what you connect: PostgreSQL, MySQL and Redshift sources are attached read-only and queried in place, and a Google Sheet is re-read from the Sheets API each session. What does persist is the answer — the rows a query returns are saved with that message in your chat history, and again in any report or dashboard you save. Files you upload are kept in full, because a file has nowhere else to live. All of it sits in your account, and you can delete it anytime. We do not use your business data to train AI models.
No credits, no question cap
Every request runs on your organization's AI provider key (your own, or on Team the owner's if you have not added one), so the model call is billed by that provider and processed under its terms — no credit balance of ours to run out on the 23rd of the month, and no usage tier to be bumped into. What leaves, precisely: your question, your table and column names, and the SQL; once a query runs, its column names, row count and first ten rows, so the written answer matches the numbers in your table (the rest of the result is not sent to the model); and, with AI Semantic Schema Learning on Pro, up to three sample rows per table when you connect or refresh a source, so the model can work out what each column means. One thing worth planning for rather than discovering: names are part of what leaves, so if a tab is called “Acme Corp 2026 orders”, rename it before you connect or withhold that table from the role you connect with.
You see the SQL that runs
You always see the exact SQL that ran. On complex questions — joins, cross-source queries, multi-step work — you get a readable query plan and approve it before anything runs; simple lookups and aggregations run straight away. When you connect a source we ask your database what that account is allowed to do and show you the answer, so you find out then rather than later whether you handed us a read-only user. On our side a keyword guard refuses write statements before any query runs — eighteen keywords matched as text, which is the second line; the read-only grant on your side is the first.
What we store
We keep what an answer produced — never a standing copy of the source it came from.
- Your account (name, email) — encrypted at rest; your password is only ever a bcrypt hash
- Your chat history — including the rows each answer returned — encrypted at rest
- The reports & dashboards you choose to save — encrypted at rest
- Your connection settings + AI key — encrypted by us with AES-256-GCM before storage
- Saved field/metric descriptions — encrypted at rest
- Files you upload (CSV/Excel) — the file itself, until you delete it; encrypted at rest
- A standing copy of your connected database or Sheet — Postgres, MySQL and Redshift are queried in place, and a Google Sheet is re-read each session
- Any row you didn't ask about — only what an answer returns is kept
- Anything used to train an AI model
The providers that process data on our behalf are named, with what each one receives, in section 5 of our Privacy Policy, which is the canonical version for a vendor review. Operated by Intellrise Technologies (SSM 202603145414), Malaysia.
Questions about security?
The one-stop AI data analyst that knows your schema.
One place from question to finished work: the SQL, the chart, a dashboard your whole team can open, and a report that emails itself every morning. It learns what your columns mean, so nobody explains a field twice. Start free with one database plus Google Sheets, or take 14 days of Pro with no card.
Every new account starts on a 14-day Pro trial with no card. After it lapses, the free tier stays free.