Data dictionary generator
Paste your SQL CREATE TABLE statements — or upload a CSV / Excel file — and get a clean, exportable data dictionary with tables, columns, types, keys, and a place to document what each field means. Free, no signup, and it runs entirely in your browser.
How it works
- Paste your DDL (
CREATE TABLEstatements) or upload a CSV / Excel file — each worksheet becomes a table. - The tool parses columns, types, nullability, and keys in your browser; for files it infers types from the sample rows.
- Fill in the business meaning for each column, then export to CSV or copy as Markdown.
Tired of maintaining a data dictionary by hand?
A dictionary goes stale the moment your schema changes. Inside Intellrise, the AI learns what your tables and columns mean from the questions you ask — and stores those definitions against the source, so everyone is working from the same stored meanings instead of a spreadsheet nobody updates. They survive later sessions and a schema refresh; the one thing that clears them is re-running a Google Sheet source's tab picker, which starts that source's schema over.
Intellrise is the one-stop AI data analyst for your own database or spreadsheet: one question becomes the SQL, the chart, a dashboard and a report, in one place. Every new account starts on a 14-day Pro trial with no card — nothing to enter, nothing to cancel — and lands on a permanently free tier after it lapses. Every plan, the trial included, needs an AI key; a Google Gemini key is free to create.
Frequently asked questions
A reference that lists every table and column in a database with its data type, keys and — most usefully — what each field actually means in business terms. It keeps a team aligned on definitions.
No. Parsing happens entirely in your browser — whether you paste SQL DDL or upload a CSV or Excel file. Your schema and data never leave your machine, and there's no account or login required.
Common CREATE TABLE syntax from PostgreSQL, MySQL/MariaDB and SQL Server — quoted identifiers, typed columns, NOT NULL, DEFAULT, inline and table-level primary and foreign keys, and MySQL COMMENT. It also reads keys attached afterwards as separate ALTER TABLE … ADD CONSTRAINT statements, which is how pg_dump and most scripted exports write them: paste the whole dump and the keys come through with the columns.
It reads the header row as column names and infers each column's type (integer, number, boolean, date or text) from the sample rows, marking a column nullable when it finds blank values. Each Excel worksheet becomes its own table. Files can't carry keys or defaults, so those are left blank for you to fill in.
That is the other direction, and it has its own tool: the CSV to SQL converter reads the same files and writes CREATE TABLE plus INSERT statements for PostgreSQL, MySQL, SQL Server, SQLite or generic ANSI SQL, with the column types inferred the same way.
Then use the ER diagram generator — it reads the same DDL with the same parser and draws it instead: a box per table, PK and FK marked, and a line for every foreign key whose target the SQL names unambiguously. It exports the SVG or a Mermaid erDiagram block. Worth knowing before you run it: a relationship your team keeps by convention rather than by a declared FOREIGN KEY gets no line there and no key marking here, for the same reason — neither tool guesses a relationship from a column name.
This free tool leaves meanings for you to type. Inside Intellrise, the AI learns what your fields mean from your questions and keeps them across sessions and through a schema refresh — so you don't maintain the dictionary by hand. (Re-running a Google Sheet source's tab picker is the one action that clears them.) If you are still choosing a tool rather than a product, we keep a side-by-side of the text-to-SQL tools that read a schema like this one, open-source options included.