How to Export Table Structure

Thank you @ScottWorld! I’ll look into these options!

I made a free tool that exports full Airtable bases (CSV + Markdown + ZIP). https://ar.konanx.com/

Didn’t work for me, at first the export was empty, but I was only trying to export the schema so that’s all that was enabled for the PAT. I added more access to the PAT and when I tried to export again there was an error about a null URL.

Can you please send some screens or videos to the DM of how to reproduce it and to see an error?

Oh yes, for sure Metadata API is what you are looking for.

If you’d like to achieve this in a no-code approach, you can look into N8Ns “native” Airtable node: Get Base Schema. -see image below. (more on n8n and other automation tools here!)


 

Mike, Consultant @ Automatic Nation

This is a script that creates a data dictionary for any base. Add it via Add an Extension, Scripting:
 

//Data dictionary for a Base

// Craig Evans datasyn.co.nz   6 June 2025

let tables = base.tables;

let dataDictionary = [];

for (let table of tables) {

    for (let field of table.fields) {

        dataDictionary.push({

            "Table": table.name,

            "Field Name": field.name,

            "Field Type": field.type,

        });

    }

}

// Sort by table then field

dataDictionary.sort((a, b) => 

    a.Table.localeCompare(b.Table) || a["Field Name"].localeCompare(b["Field Name"])

);

// Display as table in scripting extension

output.table(dataDictionary);

// Optionally: output as CSV-style string

let csv = "Table,Field Name,Field Type\n" + 

    dataDictionary.map(row => 

        `"${row.Table}","${row["Field Name"]}","${row["Field Type"]}"`).join("\n");

console.log(csv);

 For those who are still looking for a solution to this. Here’s a script you can use in the Scripting extension:

let tables = base.tables;
function getFieldNameById(fieldId) {
    if (!fieldId) return "";
    for (const tbl of tables) {
        const fld = tbl.fields.find(f => f.id === fieldId);
        if (fld) return fld.name;
    }
    return fieldId;
}
function humanizeFormula(formulaStr = "") {
    return formulaStr.replace(/\{([A-Za-z0-9]{3,})\}/g, (match, fldId) => {
        const name = getFieldNameById(fldId);
        return `{${name}}`;
    });
}
function notesForField(field) {
    const opts = field.options;
    if (!opts) return "";
    switch (field.type) {
        case "singleSelect":
        case "multipleSelects":
            return (opts.choices || []).map(c => c.name).join(", ");
        case "multipleRecordLinks": {
            const linkedTableId = opts.linkedTableId;
            const targetTable = tables.find(t => t.id === linkedTableId);
            const tableName = targetTable ? targetTable.name : linkedTableId;
            const inverseName = getFieldNameById(opts.inverseLinkFieldId);
            return opts.inverseLinkFieldId
                ? `Linked to → ${tableName} via ${inverseName}`
                : `Linked to → ${tableName}`;
        }
        case "lookup":
        case "rollup":
        case "count":
        case "multipleLookupValues": {
            const sourceName = getFieldNameById(opts.recordLinkFieldId);
            return sourceName ? `From → ${sourceName}` : "";
        }
        case "formula": {
            return humanizeFormula(opts.formula || "");
        }
        default:
            try {
                return JSON.stringify(opts);
            } catch (_err) {
                return "";
            }
    }
}
let baseSchema = [];
for (let table of tables) {
    for (let field of table.fields) {
        baseSchema.push({
            "Table": table.name,
            "Field Name": field.name,
            "Field ID": field.id,
            "Field Type": field.type,
            "Notes": notesForField(field)
        });
    }
}
baseSchema.sort((a, b) =>
    a.Table.localeCompare(b.Table) ||
    a["Field Name"].localeCompare(b["Field Name"])
);
output.table(baseSchema);