The Scenario
A platform engineer has standardized five utility functions that every RudderStack transformation should be able to import — a consistent event timestamp normalizer, a PII field scrubber, a shared error formatting helper, and two more. Each function lives in an Excel workbook: library name in column A, JavaScript code in column B. The goal is to publish all five as RudderStack shared libraries.
The engineer has been staring at this workbook for two days. The functions are done. The review passed. Now it's just the publishing step.
The bad version:
- Navigate to the RudderStack Libraries section, click "New Library," type the name from column A, paste the code from column B, click Save, click Publish.
- Go back to the list, click "New Library" again, repeat.
- Midway through library three, notice that the code block for library four didn't copy cleanly from the merged cell — spend ten minutes reformatting before continuing.
Five libraries, done carefully by hand, is a 20-minute window of repetitive UI work. The engineer was supposed to be reviewing the transformation architecture document this afternoon. This is now in the way of that.
The Easy Way: One Prompt in SheetXAI
SheetXAI is an AI agent that lives inside your Excel workbook. It reads the library spec from your workbook and creates and publishes each library in RudderStack through the API — no dashboard navigation, no row-by-row publishing.
Create a new RudderStack library for every row in my Excel table using the name from column A and the code from column B, and write each library's ID and version to columns C and D.
What You Get
- Each row becomes a published RudderStack shared library, available for import by other transformations immediately
- Column C: the library ID returned by RudderStack for each entry
- Column D: the version ID of the published version
- Any rows that fail get an error message in column C instead of an ID
- Rows where column C already has a value are left untouched
What If the Data Is Not Quite Ready
The code in column B was written with single quotes throughout
For each row where column A has a library name and column B has code, convert all single-quoted string literals in column B to double-quoted ones. Then create and publish a RudderStack transformation library using the normalized name and code, and write the library ID to column C.
Some library names have spaces — RudderStack requires camelCase names
For each row where column A has a library name and column C is blank, convert the name in column A to camelCase. Create a RudderStack library using the converted name and the code in column B, publish it, and write the library ID to column C. Write the camelCase name to column D so I can see what was actually submitted.
I want to create them as unpublished drafts first, get IDs, then publish in a second step
For each row where column A has a value and column B has code and column C is blank, create an unpublished RudderStack library draft. Write the library ID to column C and the version ID to column D. Do not publish.
Validate the function signatures, create and publish the valid ones, flag the rest, and log to a separate worksheet
For each row in Sheet1, check that the JavaScript in column B contains at least one function declaration. If it does, create and publish the RudderStack library, write the library ID to column C of Sheet1, and log the library name and ID to columns A and B of Sheet2. If the code doesn't contain a function declaration, write 'INVALID' to column C of Sheet1 and skip.
The pattern: validate, act, record — all in one instruction so the spec workbook becomes the deployment log without a separate reconciliation step.
Try It
Get the 7-day free trial of SheetXAI and open the workbook where your utility function specs are staged, then ask it to create and publish the RudderStack libraries from your name and code columns. For auditing what's already there, see Export RudderStack Library Version History Into an Excel workbook or the full RudderStack Transformation overview.