Niches · Spreadsheets & operational tables

Use a workbook's formulas to compute a result

Two or three companies' docs show a full fit.

2companies, full fit
8companies, fit with a limit
10companies carry the tag
7share of the category's work
2first-hand accounts

What counts as done

Inputs written and recalculated outputs read in the same session, with formulas evaluated by the product rather than stored as text, and the shared file left unchanged when asked.

How agents ask for it: run these inputs through our pricing model, what does the calculator give for these numbers, add a total column, recalculate the forecast

Does the job 2

The docs show it meets the test above, and nothing we read says otherwise.

ListingWhy
Grist E
Grist Labs Inc.
Python formulas are evaluated by Grist on write, and POST /docs/{id}/fork makes a personal copy to set inputs and read outputs without changing the original document. source
Microsoft Excel (Microsoft Graph workbook API) D
Microsoft Corporation
createSession with persistChanges false returns a non-persistent session, so inputs written and outputs read in that session use Excel's calculation and leave the file unchanged; /functions also exposes Excel functions directly. source

Does the job, with a limit 8

It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.

ListingWhy
Airtable BB
Formagrid Inc (doing business as Airtable)
Formula fields are computed by Airtable, so writing input fields and reading the formula field works. There is no non-persistent mode, so inputs are saved to the shared base. source
Baserow C
Baserow B.V.
Formula fields are evaluated by Baserow, so writing inputs and reading the formula column works, but there is no non-persistent mode, so the inputs change the shared table. source
Coda (Superhuman Docs) BB
Superhuman Platform Inc. (parent of Coda Project LLC)
Coda evaluates formulas, and GET /docs/{id}/formulas reads named formula results after inputs are written to rows or controls. Inputs persist to the shared doc, and writes are queued with reads from a snapshot. source
Google Sheets API A
Google LLC
valueInputOption USER_ENTERED parses formulas so Sheets evaluates them (RAW stores them as text), and a read returns the computed values. Inputs are saved to the shared file; there is no non-persistent session. source
NocoDB BB
NocoDB Inc (doing business as NocoDB)
Formula fields are computed by NocoDB and returned on read after inputs are written, but inputs persist to the shared table; there is no non-persistent mode. source
SeaTable B
SeaTable GmbH
Formula columns are computed by SeaTable and read back after inputs are written, but inputs persist to the shared base; Create Snapshot exists but there is no non-persistent session. source
Smartsheet API + MCP B
Smartsheet Inc.
A cell's formula property is evaluated by Smartsheet and read back as the computed value, with errors in the response rather than the status code. Inputs persist to the shared sheet; no non-persistent mode exists. source
Teable B
Teable AI, Inc.
Formula fields are computed by Teable and read back after inputs are written, but inputs persist to the shared table; undo exists but no non-persistent session. source

What stops agents finishing

  • formula written with RAW input stored as text
  • calculation in a shared file collides with other users

Around it

Before: spreadsheets.read-table. After: document-extraction.summarise, email.send.

Where this comes from

Weight 7 of 100 in Spreadsheets & operational tables: 2 of 35 accounts (SharePoint calculation workbooks through Graph); thin, set by hand

The same as data: /api/v1/use-cases/spreadsheets.calculate.json, or the MCP tool get_use_case with id spreadsheets.calculate. Other jobs in this category: /api/v1/niches?category=spreadsheets.

Draft, read 2026-11-15 from public docs and first-hand accounts, not yet reviewed by a person. A mark says what the docs show, not what a run measured.

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.