File emailed documents and log what they say

Whenever new mail arrives in Gmail, the workflow sweeps your inbox for messages matching a search you describe, copies the PDF attachments on each into a Box folder and adds one row per document to a register in Google Sheets. The row records when it was filed, the sender, the subject, the name it was filed under, a link to the copy and whether it could be read, followed by details read from the document itself, such as a supplier, document number, date and total, written exactly as the document gives them.

A field the document does not carry is left empty rather than guessed. Each copy keeps its own name plus a reference to its email, so this month's Invoice.pdf never collides with last month's. An email is labelled as registered only once every document on it is copied and logged. One that cannot be finished, such as an older email from a sender whose newer one is still in the inbox, is labelled 'Needs filing' for you, and any problem, including a document that could not be read, ends the run marked as failed with a note of what to do.

You need Gmail, Box, Google Sheets and OpenAI connected, a Box folder to file into, and a spreadsheet tab whose column headings match the row. Fill in the Gmail search, the Box folder and the spreadsheet link; everything else has a working default.

GmailWhen an email arrives.When an email arrives.

Set documentSearch to blank.Set documentSearch to blank.

Set boxFolder to blank.Set boxFolder to blank.

Set registerSheetUrl to blank.Set registerSheetUrl to blank.

Set sheetTabName to Documents.Set sheetTabName to Documents.

Set documentFields to Supplier, Document number, Document date, Total amount, Currency.Set documentFields to Supplier, Document number, Document date, Total amount, Currency.

Set processedLabel to Document register.Set processedLabel to Document register.

Set needsFilingLabel to Needs filing.Set needsFilingLabel to Needs filing.

Set maxDocumentsPerRun to 5.Set maxDocumentsPerRun to 5.

Set documentCap to maxDocumentsPerRun multiplied by 1.Set documentCap to maxDocumentsPerRun multiplied by 1.

If documentCap is greater than 25If documentCap is greater than 25

Yes

Stop and report a problem: The 'maxDocumentsPerRun' step needs a whole number between 1 and 25 - it is how many documents one run is allowed to handle.Stop and report a problem: The 'maxDocumentsPerRun' step needs a whole number between 1 and 25 - it is how many documents one run is allowed to handle.

No

No steps

If documentCap is less than 1If documentCap is less than 1

Yes

Stop and report a problem: The 'maxDocumentsPerRun' step needs a whole number between 1 and 25 - it is how many documents one run is allowed to handle.Stop and report a problem: The 'maxDocumentsPerRun' step needs a whole number between 1 and 25 - it is how many documents one run is allowed to handle.

No

No steps

Trim boxFolder and save it as boxFolderId.Trim boxFolder and save it as boxFolderId.

If boxFolderId contains /folder/ (ignoring case)If boxFolderId contains /folder/ (ignoring case)

Yes

Trim the first item of the last item of boxFolderId split by /folder/ split by / and save it as boxFolderId.Trim the first item of the last item of boxFolderId split by /folder/ split by / and save it as boxFolderId.

No

No steps

If boxFolderId is empty or boxFolderId multiplied by 1 is not boxFolderIdIf boxFolderId is empty or boxFolderId multiplied by 1 is not boxFolderId

Yes

Stop and report a problem: The 'boxFolder' step needs a Box folder id, which is a number - open the folder in Box and copy either its whole address or just the number after '/folder/'.Stop and report a problem: The 'boxFolder' step needs a Box folder id, which is a number - open the folder in Box and copy either its whole address or just the number after '/folder/'.

No

No steps

Set fieldNames to an empty list.Set fieldNames to an empty list.

For each rawFieldName in documentFields split by ,:For each rawFieldName in documentFields split by ,:

Trim rawFieldName and save it as fieldName.Trim rawFieldName and save it as fieldName.

If fieldName isn't emptyIf fieldName isn't empty

Yes

Add fieldName to the end of fieldNames and save it as fieldNames.Add fieldName to the end of fieldNames and save it as fieldNames.

No

No steps

If fieldNames is emptyIf fieldNames is empty

Yes

Stop and report a problem: The 'documentFields' step needs at least one field name, separated by commas - those are the columns the register records.Stop and report a problem: The 'documentFields' step needs at least one field name, separated by commas - those are the columns the register records.

No

No steps

GmailList emails matching '@documentSearch has:attachment -label:@processedLabel -label:@needsFilingLabel'. Save it as found.List emails matching '@documentSearch has:attachment -label:@processedLabel -label:@needsFilingLabel'. Save it as found.

Get the current date and time and save it as runStarted.Get the current date and time and save it as runStarted.

Make runStarted readable as “2026-11-02 14:05” and save it as filedAt.Make runStarted readable as “2026-11-02 14:05” and save it as filedAt.

Set filedCount to 0.Set filedCount to 0.

Set messagesLeft to 0.Set messagesLeft to 0.

Set setAside to an empty list.Set setAside to an empty list.

Set unreadDocuments to an empty list.Set unreadDocuments to an empty list.

Set readProblem to blank.Set readProblem to blank.

For each message in found.emails:For each message in found.emails:

If filedCount is greater than or equal to documentCapIf filedCount is greater than or equal to documentCap

Yes

Add 1 to messagesLeft.Add 1 to messagesLeft.

Skip to the next item.Skip to the next item.

No

No steps

Set messageDescription to an email.Set messageDescription to an email.

If a step below fails, save the error as messageErrorIf a step below fails, save the error as messageError

Try

Set messageId to blank.Set messageId to blank.

If a step below fails, save the error as idErrorIf a step below fails, save the error as idError

Try

Trim message.id and save it as messageId.Trim message.id and save it as messageId.

On error

No steps

If messageId is emptyIf messageId is empty

Yes

Show in the output: Skipped a message Gmail returned without an id..Show in the output: Skipped a message Gmail returned without an id..

Skip to the next item.Skip to the next item.

No

No steps

Set rawFrom to blank.Set rawFrom to blank.

If a step below fails, save the error as fromErrorIf a step below fails, save the error as fromError

Try

Trim message.from and save it as rawFrom.Trim message.from and save it as rawFrom.

On error

No steps

Set senderAddress to rawFrom.Set senderAddress to rawFrom.

If rawFrom contains < (ignoring case)If rawFrom contains < (ignoring case)

Yes

Trim the first item of the last item of rawFrom split by < split by > and save it as senderAddress.Trim the first item of the last item of rawFrom split by < split by > and save it as senderAddress.

No

No steps

Set subject to (no subject).Set subject to (no subject).

If a step below fails, save the error as subjectErrorIf a step below fails, save the error as subjectError

Try

Trim message.subject and save it as subjectText.Trim message.subject and save it as subjectText.

If subjectText isn't emptyIf subjectText isn't empty

Yes

Change subject to subjectText.Change subject to subjectText.

No

No steps

On error

No steps

Build the text '@subject' from @senderAddress and save it as messageDescription.Build the text '@subject' from @senderAddress and save it as messageDescription.

Set messageProblems to an empty list.Set messageProblems to an empty list.

If senderAddress doesn't contain @ (ignoring case)If senderAddress doesn't contain @ (ignoring case)

Yes

Add its sender could not be read (rawFrom) to the end of messageProblems and save it as messageProblems.Add its sender could not be read (rawFrom) to the end of messageProblems and save it as messageProblems.

No

Set latestId to blank.Set latestId to blank.

GmailFind email matching 'from:senderAddress'. Save it as latestFromSender.Find email matching 'from:senderAddress'. Save it as latestFromSender.

If a step below fails, save the error as latestErrorIf a step below fails, save the error as latestError

Try

Trim latestFromSender.email.id and save it as latestId.Trim latestFromSender.email.id and save it as latestId.

On error

No steps

If latestId is not messageIdIf latestId is not messageId

Yes

Add only a sender's newest email can be opened, and a newer one from this sender is in the inbox; archive that one first to the end of messageProblems and save it as messageProblems.Add only a sender's newest email can be opened, and a newer one from this sender is in the inbox; archive that one first to the end of messageProblems and save it as messageProblems.

No

GmailDownload the attachments on the latest email from senderAddress. Save it as downloaded.Download the attachments on the latest email from senderAddress. Save it as downloaded.

Set messageDocuments to 0.Set messageDocuments to 0.

Set attachmentNumber to 0.Set attachmentNumber to 0.

Set namesOnMessage to an empty list.Set namesOnMessage to an empty list.

For each attachment in downloaded.documents:For each attachment in downloaded.documents:

Set fileName to blank. Set fileName to blank.

If a step below fails, save the error as nameError If a step below fails, save the error as nameError

Try

Trim attachment.filename and save it as fileName.Trim attachment.filename and save it as fileName.

On error

No steps

Set attachmentLabel to (unnamed attachment). Set attachmentLabel to (unnamed attachment).

If fileName isn't empty If fileName isn't empty

Yes

Change attachmentLabel to fileName.Change attachmentLabel to fileName.

No

No steps

If fileName in lowercase doesn't end with .pdf (ignoring case) If fileName in lowercase doesn't end with .pdf (ignoring case)

Yes

Show in the output: Skipped attachmentLabel: only PDF attachments are treated as documents.Show in the output: Skipped attachmentLabel: only PDF attachments are treated as documents.

Skip to the next item.Skip to the next item.

No

No steps

Set documentRef to blank. Set documentRef to blank.

If a step below fails, save the error as refError If a step below fails, save the error as refError

Try

Trim attachment.document_ref and save it as documentRef.Trim attachment.document_ref and save it as documentRef.

On error

No steps

If documentRef is empty If documentRef is empty

Yes

Show in the output: Could not open attachmentLabel; it was left alone.Show in the output: Could not open attachmentLabel; it was left alone.

Skip to the next item.Skip to the next item.

No

No steps

Add 1 to attachmentNumber. Add 1 to attachmentNumber.

Replace the pattern \.pdf$ with in attachmentLabel (ignoring case) and save it as fileStem. Replace the pattern \.pdf$ with in attachmentLabel (ignoring case) and save it as fileStem.

Build the text @fileStem (@messageId).pdf and save it as boxName. Build the text @fileStem (@messageId).pdf and save it as boxName.

If namesOnMessage contains boxName If namesOnMessage contains boxName

Yes

Build the text @fileStem (@messageId @attachmentNumber).pdf and save it as boxName.Build the text @fileStem (@messageId @attachmentNumber).pdf and save it as boxName.

No

No steps

Add boxName to the end of namesOnMessage and save it as namesOnMessage. Add boxName to the end of namesOnMessage and save it as namesOnMessage.

Set storedLink to blank. Set storedLink to blank.

Set uploadProblem to blank. Set uploadProblem to blank.

If a step below fails, save the error as uploadError If a step below fails, save the error as uploadError

Try

BoxUpload boxFolderId to Box folder documentRef. Save it as stored.Upload boxFolderId to Box folder documentRef. Save it as stored.

Trim stored.file_link and save it as storedLink.Trim stored.file_link and save it as storedLink.

On error

Split uploadError.message by a new line and save it as uploadErrorLines.Split uploadError.message by a new line and save it as uploadErrorLines.

Get the first item of uploadErrorLines and save it as uploadProblem.Get the first item of uploadErrorLines and save it as uploadProblem.

If a step below fails, save the error as lookupErrorIf a step below fails, save the error as lookupError

Try

BoxList items in Box folder boxFolderId. Save it as folderItems.List items in Box folder boxFolderId. Save it as folderItems.

Keep each item in folderItems.items where item.name is boxName and save it as sameName.Keep each item in folderItems.items where item.name is boxName and save it as sameName.

If sameName isn't emptyIf sameName isn't empty

Yes

Build the text https://app.box.com/file/@sameName[0].id and save it as storedLink.Build the text https://app.box.com/file/@sameName[0].id and save it as storedLink.

No

No steps

On error

No steps

If storedLink is empty If storedLink is empty

Yes

If filedCount is 0If filedCount is 0

Yes

Stop and report a problem: Nothing was filed this run: attachmentLabel on messageDescription could not be copied into Box (uploadProblem). Check that Box is connected and that the 'boxFolder' step names a folder you can add files to. The email is left as it was, so the next run tries it again.Stop and report a problem: Nothing was filed this run: attachmentLabel on messageDescription could not be copied into Box (uploadProblem). Check that Box is connected and that the 'boxFolder' step names a folder you can add files to. The email is left as it was, so the next run tries it again.

No

No steps

Add attachmentLabel could not be copied into Box (uploadProblem) to the end of messageProblems and save it as messageProblems.Add attachmentLabel could not be copied into Box (uploadProblem) to the end of messageProblems and save it as messageProblems.

Skip to the next item.Skip to the next item.

No

No steps

Set readStatus to Read. Set readStatus to Read.

Set returnedNames to an empty list. Set returnedNames to an empty list.

Set returnedValues to an empty list. Set returnedValues to an empty list.

If a step below fails, save the error as extractError If a step below fails, save the error as extractError

Try

Run an AI agent: Read the attached document and pull out each of these fields, in this order: - fieldName… and save it as extraction.Run an AI agent: Read the attached document and pull out each of these fields, in this order: - fieldName… and save it as extraction.

For each returnedField in extraction.output.fields:For each returnedField in extraction.output.fields:

Set returnedName to blank. Set returnedName to blank.

If a step below fails, save the error as fieldNameError If a step below fails, save the error as fieldNameError

Try

Trim returnedField.name and save it as returnedName.Trim returnedField.name and save it as returnedName.

On error

No steps

Set returnedValue to blank. Set returnedValue to blank.

If a step below fails, save the error as fieldValueError If a step below fails, save the error as fieldValueError

Try

Trim returnedField.value and save it as returnedValue.Trim returnedField.value and save it as returnedValue.

On error

If a step below fails, save the error as fieldListErrorIf a step below fails, save the error as fieldListError

Try

Join returnedField.value with , and save it as returnedValue.Join returnedField.value with , and save it as returnedValue.

On error

No steps

Add returnedName to the end of returnedNames and save it as returnedNames. Add returnedName to the end of returnedNames and save it as returnedNames.

Add returnedValue to the end of returnedValues and save it as returnedValues. Add returnedValue to the end of returnedValues and save it as returnedValues.

On error

Change readStatus to Filed but could not be read.Change readStatus to Filed but could not be read.

Add boxName to the end of unreadDocuments and save it as unreadDocuments.Add boxName to the end of unreadDocuments and save it as unreadDocuments.

Change readProblem to extractError.message.Change readProblem to extractError.message.

Set rowValues to an empty list. Set rowValues to an empty list.

Add filedAt to the end of rowValues and save it as rowValues. Add filedAt to the end of rowValues and save it as rowValues.

Add senderAddress to the end of rowValues and save it as rowValues. Add senderAddress to the end of rowValues and save it as rowValues.

Add subject to the end of rowValues and save it as rowValues. Add subject to the end of rowValues and save it as rowValues.

Add boxName to the end of rowValues and save it as rowValues. Add boxName to the end of rowValues and save it as rowValues.

Add storedLink to the end of rowValues and save it as rowValues. Add storedLink to the end of rowValues and save it as rowValues.

Add readStatus to the end of rowValues and save it as rowValues. Add readStatus to the end of rowValues and save it as rowValues.

For each wantedField in fieldNames: For each wantedField in fieldNames:

Set cellValue to blank. Set cellValue to blank.

Find the position of wantedField in returnedNames and save it as foundAt. Find the position of wantedField in returnedNames and save it as foundAt.

If foundAt is greater than or equal to 0 If foundAt is greater than or equal to 0

Yes

Change cellValue to returnedValues[@foundAt].Change cellValue to returnedValues[@foundAt].

No

No steps

Add cellValue to the end of rowValues and save it as rowValues. Add cellValue to the end of rowValues and save it as rowValues.

If a step below fails, save the error as registerError If a step below fails, save the error as registerError

Try

Google SheetsAdd a row to sheetTabName in registerSheetUrl.Add a row to sheetTabName in registerSheetUrl.

On error

If filedCount is 0If filedCount is 0

Yes

Stop and report a problem: Nothing was added to the register this run: the row for boxName could not be written (registerError.message). Check that Google Sheets is connected and that 'registerSheetUrl' and 'sheetTabName' point at your register. The copy is already in Box; the email is left as it was, so the next run tries again and uses that copy.Stop and report a problem: Nothing was added to the register this run: the row for boxName could not be written (registerError.message). Check that Google Sheets is connected and that 'registerSheetUrl' and 'sheetTabName' point at your register. The copy is already in Box; the email is left as it was, so the next run tries again and uses that copy.

No

No steps

Add boxName is in Box but could not be added to the register (registerError.message) to the end of messageProblems and save it as messageProblems.Add boxName is in Box but could not be added to the register (registerError.message) to the end of messageProblems and save it as messageProblems.

Skip to the next item.Skip to the next item.

Add 1 to filedCount. Add 1 to filedCount.

Add 1 to messageDocuments. Add 1 to messageDocuments.

Show in the output: Filed boxName and added it to the register. Show in the output: Filed boxName and added it to the register.

If messageProblems is empty and messageDocuments is 0If messageProblems is empty and messageDocuments is 0

Yes

Show in the output: Nothing to register on messageDescription - it carried no PDF.Show in the output: Nothing to register on messageDescription - it carried no PDF.

No

No steps

If messageProblems is emptyIf messageProblems is empty

Yes

GmailAdd label processedLabel to email messageId.Add label processedLabel to email messageId.

No

Join messageProblems with ; and save it as problemText.Join messageProblems with ; and save it as problemText.

Add messageDescription: problemText to the end of setAside and save it as setAside.Add messageDescription: problemText to the end of setAside and save it as setAside.

GmailAdd label needsFilingLabel to email messageId.Add label needsFilingLabel to email messageId.

On error

If filedCount is 0If filedCount is 0

Yes

Stop and report a problem: Nothing was filed this run: messageDescription could not be handled (messageError.message). The email is left as it was, so the next run tries it again.Stop and report a problem: Nothing was filed this run: messageDescription could not be handled (messageError.message). The email is left as it was, so the next run tries it again.

No

No steps

Add messageDescription: messageError.message to the end of setAside and save it as setAside.Add messageDescription: messageError.message to the end of setAside and save it as setAside.

If a step below fails, save the error as setAsideErrorIf a step below fails, save the error as setAsideError

Try

GmailAdd label needsFilingLabel to email message.Add label needsFilingLabel to email message.

On error

No steps

If filedCount is 0 and setAside is empty and messagesLeft is 0If filedCount is 0 and setAside is empty and messagesLeft is 0

Yes

Finish the run: No new documents to add to the register.Finish the run: No new documents to add to the register.

No

No steps

Set reportLines to an empty list.Set reportLines to an empty list.

Add Filed filedCount document(s) into Box and onto the register. to the end of reportLines and save it as reportLines.Add Filed filedCount document(s) into Box and onto the register. to the end of reportLines and save it as reportLines.

If messagesLeft is greater than 0If messagesLeft is greater than 0

Yes

Add messagesLeft more message(s) are waiting: this run reached its limit of maxDocumentsPerRun documents, and the next run carries on from there. to the end of reportLines and save it as reportLines.Add messagesLeft more message(s) are waiting: this run reached its limit of maxDocumentsPerRun documents, and the next run carries on from there. to the end of reportLines and save it as reportLines.

No

No steps

If unreadDocuments isn't emptyIf unreadDocuments isn't empty

Yes

Join unreadDocuments with , and save it as unreadList.Join unreadDocuments with , and save it as unreadList.

Add Filed, but the details could not be read, so their rows are empty for you to fill in: unreadList. The reader said: readProblem to the end of reportLines and save it as reportLines.Add Filed, but the details could not be read, so their rows are empty for you to fill in: unreadList. The reader said: readProblem to the end of reportLines and save it as reportLines.

No

No steps

If setAside isn't emptyIf setAside isn't empty

Yes

Join setAside with "Join setAside with - and save it as setAsideList.

Add These emails could not be finished and are labelled '@needsFilingLabel'. Fix the cause and remove the label to have one tried again, or file it by hand:Add These emails could not be finished and are labelled '@needsFilingLabel'. Fix the cause and remove the label to have one tried again, or file it by hand: - setAsideList to the end of reportLines and save it as reportLines.

No

No steps

Join reportLines with a new line and save it as runReport.Join reportLines with a new line and save it as runReport.

If setAside isn't emptyIf setAside isn't empty

Yes

Stop and report a problem: runReport.Stop and report a problem: runReport.

No

No steps

If unreadDocuments isn't emptyIf unreadDocuments isn't empty

Yes

Stop and report a problem: runReport.Stop and report a problem: runReport.

No

No steps

Finish the run: runReport.Finish the run: runReport.