XLS Tools
Utilities & PersonalCreate and read spreadsheet files inside FlowRunner with no API key, vendor account, or outbound call beyond fetching the file itself. Agents build and parse workbooks as a native workflow step.
What This Integration Enables
XLS Tools creates and reads spreadsheet files inside FlowRunner. There is no API key and no vendor account, and no outbound call beyond fetching the file you point it at, so a flow that handles the workbooks a business actually exchanges carries no credential and no third-party service in the path. Spreadsheets remain the real interchange format for most organisations, whatever the system of record says, and a platform that cannot open one has a gap exactly where the work is. The connector builds .xlsx workbooks from flow data with native types preserved, including multi-sheet workbooks in a single pass, parses uploaded workbooks into rows, reports a workbook's structure without returning any cell data, and reads and writes RFC 4180 delimited text with full control over delimiter, quoting, line endings and encoding. It is also careful in the places spreadsheets are dangerous: formula escaping is on by default when writing delimited text, so a workbook from an untrusted upload cannot produce a file that executes a formula when a colleague opens it, and parsed values stay as text unless conversion is explicitly requested, which is what protects leading zeros in postcodes and account numbers.
Without FlowRunner
With FlowRunner
Use Case Scenarios
The emailed attachment becomes rows
A supplier sends a remittance advice as an .xlsx attachment every month. An agent watching the mailbox calls Get Workbook Info first, which returns the worksheet names in order with each one's row and column count and the workbook timestamps, without returning a single cell. That is enough to branch the flow on which tabs exist and to reject an obviously wrong upload before a heavy parse. The agent then calls Parse Excel File on the expected sheet, receives rows as objects keyed by the header row, and posts them into QuickBooks Online as draft entries. Duplicate header names are made unique with a numeric suffix, so a workbook with two columns called Amount loses neither of them.
A report a colleague can open
A weekly reporting flow queries Snowflake for regional performance and needs to put it in front of people who will not log into a BI tool. The agent calls Create Multi-Sheet Excel File with one worksheet per region, each with its own column order, and gets back a single downloadable workbook. Numbers stay numeric and booleans stay boolean, so the recipients can sort and total the columns without cleaning them first. The file link is posted to the leadership channel in Slack with the summary in the message body.
A legacy export turned into something usable
A warehouse system produces a semicolon-delimited export with a byte order mark and no consistent quoting. The agent calls Convert CSV To Excel, which detects the delimiter by comparing column consistency across the first lines, strips the byte order mark, and converts numeric and true or false text into real Excel types so the resulting columns total correctly. Nothing in the file is interpreted as a formula on the way in. The workbook is filed in Microsoft OneDrive where the operations team already keeps its records.
Human-in-Loop Highlight
The gate is the parse that quietly succeeded on the wrong file. A recurring workbook is the most reliable input a finance flow has, right up until the sender inserts a column, renames a tab or ships a partial month, at which point the parse still returns rows and the import still succeeds and the numbers are simply wrong in a way that surfaces weeks later in a reconciliation. Get Workbook Info exists precisely so an agent can check before it commits: worksheet names in order, row and column counts, and workbook timestamps, with no cell data returned. So before any parsed workbook reaches a system of record, the agent compares that structure against the shape it accepted last time, and when the tab list, the header set or the row count moves outside the tolerance, it stops and puts the difference in front of the controller in Microsoft Teams: "Kestrel remittance for August has 3 tabs instead of 2, the Detail sheet gained a column called FX Adj, and the row count is 118 against 340 last month. Import anyway, import the matching columns only, or hold and query with the supplier?" The agent could have imported it. The point of human-in-the-loop is that it noticed it should not, and asked the one person who knows whether a partial month is expected.
Agent Capabilities
10 actionsExcel Workbooks
4- Create Excel File Builds an .xlsx workbook from a list of rows and stores it in FlowRunner file storage, returning a downloadable URL. Rows may be objects keyed by column name, in which case the header row is generated from the keys, or plain arrays of values, and an explicit column list fixes the order or exports a subset. Values are written with their native type, so numbers stay numeric and booleans stay boolean, and strings beginning with an equals sign can optionally be written as live formulas. Only the modern .xlsx format is produced.
- Create Multi-Sheet Excel File Builds an .xlsx workbook containing several worksheets in one pass, each with its own name, rows and optional column order. Used to export related datasets, such as orders and line items or one tab per region, into a single deliverable file rather than a folder of attachments.
- Parse Excel File Downloads an .xlsx workbook and returns one worksheet as structured rows, either as objects keyed by the header row or as arrays of raw values. Formula cells return their last calculated result, rich text is flattened, hyperlinks return their display text and dates come back as ISO 8601 strings. Duplicate header names are made unique with a numeric suffix so no column is silently lost. Legacy .xls, .xlsb and .ods files are rejected with a clear error rather than partially read.
- Get Workbook Info Downloads an .xlsx workbook and reports its structure without returning any cell data: worksheet names in order, the row and column count of each, and the workbook creation and modification timestamps. Used to branch a flow on which tabs exist, to validate an upload before a heavy parse, and to detect that a recurring file has changed shape.
Delimited Text
4- Create CSV File Builds an RFC 4180 delimited text file from a list of rows and stores it in FlowRunner file storage. The delimiter, line ending and quoting aggressiveness are all configurable, and a UTF-8 byte order mark can be prepended so Excel opens accented characters correctly. Formula escaping is on by default, which prefixes any value starting with a formula-triggering character so the file cannot execute anything when opened.
- Create CSV Text Renders rows as delimited text and returns it as a string rather than writing a file, which suits small payloads going straight into an email body, an API request or a code step. Supports the same delimiter, quoting, line ending and formula escaping choices. Large exports belong in Create CSV File, because the whole document is held in memory and returned inside the flow payload.
- Parse CSV File Downloads a delimited text file and returns its records as objects keyed by the header row or as arrays of raw values, following RFC 4180 so quoted fields may contain the delimiter, embedded line breaks and doubled quotes. The delimiter can be detected automatically, a byte order mark is stripped, duplicate headers are made unique and non-UTF-8 files can be read by selecting the encoding. Values stay as text unless numeric and boolean conversion is enabled, which is what protects leading zeros in postcodes and account numbers.
- Parse CSV Text Parses delimited text already in the flow, whether pasted into a field, pulled from an API response or read out of an email body, with no file download. Follows the same RFC 4180 rules and automatic delimiter detection, returning records as objects or arrays with optional type conversion.
Conversion
2- Convert Excel To CSV Downloads an .xlsx workbook, reads one worksheet and writes it back out as a delimited text file in FlowRunner file storage. Formula cells export their last calculated result, dates become ISO 8601 strings and rich text is flattened. Formula escaping is on by default so a workbook from an untrusted upload cannot produce a file that executes on open. This is the fastest route from an Excel attachment to a system that only accepts delimited text.
- Convert CSV To Excel Downloads a delimited text file and writes it back out as an .xlsx workbook. The delimiter can be detected automatically, a byte order mark is stripped, and numeric and boolean text can be converted into real Excel types so the resulting columns sort and total correctly. Every value is written as literal text or a number, and nothing is interpreted as a formula.
Frequently Asked Questions
What can FlowRunner do with XLS Tools?
FlowRunner agents can run Create Excel File, Create Multi-Sheet Excel File, and Parse Excel File in XLS Tools, plus 7 more actions.
Does connecting XLS Tools to FlowRunner require OAuth?
No. XLS Tools needs no credentials at all, so there is no key to obtain, store, or rotate.
Can XLS Tools trigger a FlowRunner workflow automatically?
XLS Tools doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with XLS Tools
$100 in credits. No card required. Connect in minutes.