Files

A library of every file uploaded against a task or submission, and a search that finds words inside files.

The Files page has two tabs:

• All files is a read-only library that pulls every file attached to a task or submitted against a task into one searchable list. You manage files at their source; this tab is for finding them.
• Search contents finds words written inside files: the text of a document, a scanned page, a picture, or what was said in a recording. It looks across every place your files live (chats, WhatsApp conversations, tasks, expenses, lead calls, clients and hiring), and only shows the files you are allowed to open.

Where it lives

Workspace, Files. The tabs are at the top of the page, under the title.

All files

Files in this list come from two places: task attachments (one per task) and task submissions (multiple per task).

Finding a file

The toolbar at the top has:

• Search. Matches against the filename, the task title, and the name of the person who uploaded it. Filters live as you type. (To search the words inside the files, use the Search contents tab.)
• Source filter. Tick Task attachments, Submissions, or both.
• Type filter. Tick Files, Links, or both.

Active filters appear as pills under the toolbar. Click a pill's X to remove that filter, or Clear all to reset.

Reading a row

Each file shows:

• An icon that tells you where it lives:
  • Google Drive logo for Google Drive files
  • A cloud icon for files in managed cloud storage
  • A generic file or link icon for everything else
• Filename, clickable. Opens the actual file in a new tab.
• Task title, clickable. Jumps to the task's detail page.
• Source badge: Attachment or Submission.
• Team badge (if the task has a team), coloured to match the team.
• Extracted text button, for a file in the app's cloud storage that the app has read, is reading, or couldn't read (the viewer says which; see below). It is hidden while the workspace's subscription is suspended.
• Uploader: avatar initials, name, and how long ago they uploaded.

Search contents

Open the Search contents tab and type at least 2 characters (up to 200). The search runs a moment after you stop typing, and the previous results stay on screen, dimmed, while the next ones load.

It looks at:

• the words inside the file: a document's text, the text of scanned pages and pictures, a recording's transcript;
• the file's name, and its own title if it has one;
• a few details the file carries, such as a document's author, subject and keywords, an email's sender and recipients, a spreadsheet's sheet names, or the make and model of a camera.

Searching in Hinglish or in your own script

You can type Hindi, Marathi or another Indian language in English letters (often called Hinglish), for example kya ye hindi, or in the language's own script, for example क्या यह हिंदी. The tab finds the words written either way, in a document, a scanned page, a recording's transcript or a file's name: a file in the language's script is found by its spelling in English letters, and the reverse. Spellings that people write in different ways find each other too, such as kam and kaam, or nahi and nhi.

A match found this way is highlighted in the words as they were written in the file, not as you typed them. Matches exactly as you typed them come first; a file found only through another spelling or script comes after them. While you type, the last, unfinished word already matches: kya y finds क्या यह. A search of several words looks for files with all of the words close together (in the same passage; a file's title counts as part of every passage) or all in the file's name; if there is none, a search of three or more words tries again without its last words. In a file's name, only whole words match: finish typing a word to find a file by its name (what a file says already matches while you type). A file named in Chinese or Japanese characters is found by what it says rather than by its name, and so are words that a name joins with a symbol such as € or |, and Arabic words that start with "ال" or hold "ي", "ى" or "ك". In a Korean, Thai or Khmer name, only a whole run of letters between spaces or most punctuation marks matches, never the start of one.

A search with a word that starts with a minus sign (budget -draft) or a phrase in "quotes" matches only the words as typed, with no spelling matching. Not every spelling is covered: if a word doesn't turn up, try another common spelling of it, or write it in the language's own script. Languages written in the Arabic script, such as Urdu, are found only as typed.

Reading a result

Results are grouped by file. Each one shows:

• The file name, with the words that matched highlighted, and what kind of file it is (PDF, Document, Spreadsheet, Presentation, Image, Audio, Video, Text file, Archive, Email), plus the document's own title when it differs from the name.
• Up to three matches: the words around each match, under a small label that says where in the file it is: Page 3, Sheet: Budget, Slide 4, a time in a recording such as 0:42, or Details for a match in the file's details.
• In, followed by links to the places you can open the file: Team chat: Design, Task: Ship v2, Expense: Flight to Pune, Lead call: Asha Rao, Client: Acme, Job application: Asha Rao, Designer, and so on. A call's recording also shows when the call was. If you can open the file in more places than are listed, the line ends with and N more places.
• Open file (opens the file itself) and Extracted text (opens the viewer described below).

When a match has a place label (a page, a sheet, a slide or a time), click it and the extracted text opens right there, with that part highlighted for a moment.

A search shows up to 20 files. If more match, a note under the list says so: add more words to narrow it down.

If no file matches, the tab says nothing was found in the files read so far. A file becomes searchable once its text has been read, so new (or older) files may not show up yet. If files aren't read on your deployment at all, the tab says Searching inside files is off (you can still search file names on All files). If the search can't run for a moment, the tab says so and offers Retry, rather than telling you nothing matched.

Extracted text

Extracted text opens a viewer with everything that was read out of one file:

• A document, spreadsheet, presentation, scan or picture shows its text under a heading for each part: Page 1, Sheet: Budget, Slide 4, Image 2, or an attachment of an email.
• An audio or video file shows its transcript: each line with the time it starts, who spoke when the voices could be told apart (Speaker 1, Speaker 2), and what was said.
• About this text lists, in plain words, anything that was left out: a password-protected file, scanned pages or pictures that couldn't be read, only the beginning of a long recording, and similar.
• Open file in the footer opens the original.

The viewer shows up to the first 100,000 characters of a file's text and says so when there is more. A file that isn't ready says why: This file is still being read. Its text usually appears within a few minutes. (a file stored recently that can be read now), Transcribing… (a recording on its way), This file hasn't been read yet. (it can't be read right now, or has waited a long while, so no time is promised), This file's text couldn't be read, This file is too large to read, This file's text was not read, No text was found in this file, or, for a recording, No speech detected or This recording has no transcript. An open viewer checks again by itself until the file is read.

What can be read

Text in other scripts (for example Arabic, Chinese or Cyrillic) isn't read from pictures and scans, and the app doesn't read the picture itself, only the text in it. Files that live in Google Drive, and links, are not read: the app doesn't hold their contents. Only files in the app's own cloud storage are.

When a file is read

A file is read in the background shortly after it is stored: usually within a few minutes, longer for a large scan or a long recording. You don't need to do anything. Until then, Search contents doesn't find the file at all (not even by its name): use the All files tab's search to find it by name. Files stored before this was switched on for your deployment are read only once your platform administrator processes them, so they may take a while to become searchable, or may not be searchable yet.

Limits to know about:

• Long recordings. Only the first part of an audio or video file is transcribed (the first 10 minutes, unless your platform administrator has changed it). The viewer says how much was transcribed.
• Long scans. By default the first 50 scanned pages of a document are read; the viewer says when pages were left out.
• Large files. A file over 200 MB isn't read at all, and from a long document only the first part of its text is searchable (the first 150,000 characters).

To transcribe a recording, its sound is sent securely to a speech-to-text service. The text of documents, scans and pictures is read on the platform's own servers and never leaves them.

Who sees what

The Search contents tab only ever lists files you could open yourself on the page that shows them, and it checks this on every search. Someone who leaves the workspace, or loses a permission, stops finding those files at once.

Files from chats, WhatsApp, expenses, leads, clients and hiring are only read in workspaces that have those features. A file that appears in more than one place is shown through the places you may open. The places you can't open are never listed.

In a workspace whose subscription is suspended, both the search and the extracted text are paused; All files still lists the files.

Opening a file from its link

A file's link only opens for people who could open the place it belongs to, the same people the search lists it for. Having the link is not enough: a colleague who is not in a direct message, cannot see a client, cannot open the task, or lacks the permission for that area gets "No access to this file".

Uploads follow the same idea in reverse: you can add a file to a folder only if you have the permission for it (task files need a task permission, receipts an expense permission, a client's files Manage Clients on a client you can open, short link images a short link permission).

Deleting or replacing a file

You can't delete from this page. To replace a task's attachment, open the task and edit the attachment field. To remove a submission file, open the task and delete the submission. A file that was replaced or deleted stops showing up in searches.

Cloud files vs Google Drive

Depending on which integrations are enabled for your organisation, files land in different places:

• Cloud files. The default whenever your org has cloud storage enabled, even for people who have linked a Google account. Uploads go to the app's managed cloud storage, and everyone who may open the place the file belongs to can open it.
• Google Drive. Pick it in the upload control if you have linked a Google account. The file goes to your own Google Drive and stays private to you: it is not shared with "anyone with the link", so colleagues can't open the stored link until you share the file from Drive. Choose Cloud for anything your team needs to open. When Drive is the only option, the control says the file is saved privately. If Drive can't take the upload, you get a plain message and can try again or choose Cloud.

Either way the file shows up in the All files library, you just see a different icon. Only files in cloud storage can be read for text, and files kept in Google Drive don't count toward the workspace's storage limit.

Permissions and visibility

• Files. Opens the Files page. The All files tab lists the attachments and submitted files of the tasks you can see.
• View All Files. See every task attachment and submitted file in the org, including on tasks you can't open.
• Without it: you only see files on tasks you created, tasks you're assigned to, tasks belonging to a team you're a lead of, or unclaimed tasks of your teams. A file someone else submitted is listed only to the person who submitted it and the task's creator.
• Admin and Manager roles see everything by default. Members have the Files permission only, so they see the files on their own tasks.
• Search contents follows the rules in Who sees what, above.