Connect your knowledgeUpload content

Upload content

Upload files to Gurusup before and after your first build. See the limits, every refusal message with its fix, and how to check that your files arrived.

Upload files to teach Gurusup what does not live in a connected app: policies, decks, spreadsheets, exports, contracts.

For admins and members. Both can upload files and import a wiki from Sources. Setup, Ingest, webhook sources and connecting apps are for admins.

Before you start

  • You are an admin or a member of the organization, with active access.
  • Ingestion is running. In Demo Mode, or while ingestion is paused, uploading is off. See Check source status.
  • Your files are supported. See What you can upload and Limits.

Where you upload depends on your first build

Your memoryWhere you uploadCan you remove a file afterwards?
Not built yetSetup, step Connect your sources, or Home, through Drop files or Migrate your current wikiYes, until you build.
Building or builtUpload files on the Sources pageNo. See Replace or remove a file.

Until the first build, Sources and Ingest only show a preview, so Setup and Home, which are for admins, are the only places to upload. Members upload from Sources once the memory is built.

Upload during setup

Use the setup step Connect your sources, or Home while it reads No sources connected yet. After you choose Connect live sources, the Drop files card sits under Additional sources. To move a whole wiki instead, see Import a whole wiki.

Open the upload panel

On the Drop files card, click +. A side panel opens, titled Upload documents.

Add your files

Drag files into the drop area, or click it to browse. You can add several at once, up to 256 MiB each. Each file gets a row under Uploaded files and uploads on its own. There is no usage estimate on this screen.

Wait for the rows to settle

You should see "Ready for your first build" on each row. If a row shows a reason in red instead, see What the rows show.

Close the panel and build

Click Done. The card turns into a green check labeled Manage Drop files. Click Continue in setup. The button stays off while a file is still uploading. After setup, Gurusup starts the build by itself once a document is ready. Building is covered in the Quickstart.

Files you upload here are first in line for the first build, and they get Default access: the setup panel has no access choice.

What the rows show

Each row shows the file name, a status line and its size in KB.

You seeWhat it meansWhat to do
"Waiting to upload…" or "Uploading…"The file is queued, uploading or being accepted.Wait. Continue stays off until every upload is done.
"Preparing file…"Gurusup accepted the file and is getting it ready.Wait.
"Ready for your first build"The file is ready.Nothing. Click Remove if you added it by mistake.
"Uploaded"The file is ready and can no longer be removed.Nothing.
A status saying the file is waiting for available usageYour usage is spent.See Plan and usage.
"Could not prepare this file. Remove it and upload it again."Gurusup could not get the file ready.Click Remove, then upload it again.
A reason in red under the file name, with Retry or Choose another file, and RemoveThe upload or the check of that file failed. The other files carry on.Read the reason and look it up in Refusals and fixes. Click Retry when the problem is temporary, or Choose another file when Gurusup refused the file. The new file takes the place of the old row. Remove takes the row out.
"Demo Mode is read-only. You can view your uploaded files."Demo Mode is on. The drop area is off.Nothing to do in the app.
"Ingestion is paused. You can view or remove uploaded files. Ask your organization admin to resume ingestion to add files."Ingestion is paused. The drop area is off, and Remove still works.Remove a file if you need to. To add files, wait until ingestion resumes.

Upload from Sources

Use this once your memory is building or built.

Open the dialog

Go to Sources and click Upload files. The Upload files dialog opens. The line under the drop area lists the accepted formats and "Up to 256 MiB per file."

Add your files

Drop files anywhere on the dialog, or click the drop area. Under it, the dialog reads "Documents, spreadsheets, presentations, PDFs, text files and wiki exports, or a ZIP of them." and the size limit per file. See supported files lists every extension. Each file appears under Selected files while Gurusup uploads it and works out its usage. A file that cannot be prepared shows its reason on its own row and does not stop the others: a file over the size limit no longer clears the whole selection. You cannot add or remove files while Gurusup is still preparing the current ones.

Choose who reads them

Under Knowledge access, choose Default access, Restricted or Private. After your first upload, the dialog shows the current access of Local Uploads and says "Applies to every uploaded file, including the ones already uploaded." Change it only if you mean to. Members see a shorter choice, without that line. See Choose who reads your files.

Read the estimate

Each file shows its estimated usage. Under the list you see "Estimated usage: ... for this upload · N% left this period". A file with an error shows no estimate, and its reason sits under its name. Inside an archive, a file Gurusup cannot read is not counted: the dialog says "N files can't be read and won't be added" and lists a reason for each, naming the file and what to do. Those files are skipped.

Confirm

Click Confirm upload. If you changed Local Uploads access, Gurusup saves it before uploading the files. If the save fails, no files upload and the dialog keeps your selection. If the upload takes more than 10% of your usage for the period, a summary comes first: click Confirm upload again, or Back. You should see "Accepted N files, queued for processing. Processing continues in the background." Files that could not be prepared stay in the dialog, because the dialog reads "Files that could not be prepared won't be added. Confirm upload adds only the ready files." If Gurusup skipped files it adds which ones, for example "a.key wasn't added: Gurusup can't read that kind of file. Save it as a PDF, Word or Excel file and add it again." For several files it reads "2 files weren't added because Gurusup can't read them" followed by the first names. The dialog closes and you can leave the page.

Closing the dialog or clicking Cancel before you confirm discards the whole selection, and Gurusup deletes the temporary copies. Nothing is added to your memory until you confirm.

Choose who reads your files

Every file an admin uploads goes to one source, Local Uploads, so one access setting covers all of them.

Knowledge accessWho reads it in Brain
Default accessEveryone, according to what they are already allowed to see: admins every category, members and agent accounts the categories an admin granted them.
RestrictedOnly the people you pick. Administrators are not included unless you pick them.
PrivateOnly you. Administrators are not included.
WhenWhat you can choose
Your first upload from SourcesChoose Default access, Restricted or Private. The settings apply to Local Uploads, including future files.
Later uploadsEdit Local Uploads access in the dialog. It applies to every uploaded file, including files already uploaded. Changes save when you click Confirm upload, before Gurusup uploads files. If saving fails, no files upload.
Uploads in setupNothing. They get Default access.

If some files must stay with one team, do not upload them: send them to a webhook source created and restricted for that team. See Restrict a source to specific people.

If you are a member

The dialog says "Choose who can read these files. Files you uploaded before keep their access." and offers Default access, Restricted and Private. Your files go to a source of their own for each choice, grouped under Local Uploads, so a new choice never changes who reads files you uploaded before. For the three options, see Restrict a source to specific people.

Under Restricted, you are always included. If nobody else is in the organization yet, Gurusup says "No teammates in your organization yet." If the list of people does not load, Gurusup says "Could not load members. Keep the files private or try loading the list again." Click Load members, or choose Private.

If Confirm upload stays off

CauseWhat to do
Gurusup is still preparing a fileWait for the estimates.
Gurusup is still loading your sources, to know whether Local Uploads existsWait a moment.
A row shows a reason in redClick Retry or Choose another file on that row, or remove it (the X). The dialog says "Files that could not be prepared won't be added. Confirm upload adds only the ready files.", so the other files can still go.
Nothing in the selection can be usedAdd a supported file.
"Resolve colliding Source Item Paths before uploading:" with a count for each pathTwo entries share a path. Remove or rename one.

Import a whole wiki

Pick the tool your wiki lives in. Gurusup shows that tool's export steps above the usual upload area, and each page or note comes in as its own document. Formats, limits and refusals are the same as for any upload.

WhenWhere
In setup, or on Home before the first buildMigrate your current wiki (on Home, click All options first), then + on your tool's card. The panel is titled Import from and the tool, or Import your wiki for Other. It works like Upload during setup.
Once the build is confirmedSources, then Import wiki, which opens Import from another wiki. Pick a tool and the Upload files dialog opens, titled Import from and the tool. It works like Upload from Sources.
ToolHow to export itWhat you drop
ConfluenceOpen the space, then Space settings → Export space. Choose HTML and export the whole space.The ZIP
ObsidianCompress your whole vault folder into a ZIP. Canvases come in too. Obsidian's settings and trash are left out.The ZIP
LogseqCompress your graph folder, the one with pages and journals inside, into a ZIP. Block references are kept. Logseq's settings are left out.The ZIP
EvernoteRight-click a notebook (or select notes), choose Export and pick ENEX. Attachments are named in each note but not read.The .enex file
NotionOpen Settings → Workspace → General and choose Export all workspace content. Pick Markdown & CSV and turn on Include subpages. To keep it up to date instead, connect Notion.The ZIP Notion sends you
BookStackOpen the book (or chapter), then Export → Portable ZIP.The ZIP
MediaWikiOpen Special:Export on your wiki, list the pages or add a category, and keep Include only the current revision. Templates and redirects are left out.The XML file
Roam ResearchOpen ••• in the top right, then Export All, and pick JSON. Block references are kept.The JSON file
TiddlyWikiExport your tiddlers as JSON (Tools → Export all). System tiddlers and images are left out.The JSON file
GitBookSync the space with GitHub or GitLab, then download the repository as a ZIP.The ZIP
OutlineOpen Settings → Export and export all collections as Markdown.The ZIP
JoplinChoose File → Export all → MD - Markdown, then compress the exported folder into a ZIP.The ZIP
Markdown folderCompress the folder into a ZIP, including a wiki your AI agent keeps. Folders and links are kept. An agent's CLAUDE.md or AGENTS.md is left out.The ZIP
OtherBear or Ulysses: export notes as TextBundle. Tettra, Document360, Slite or another wiki: export as HTML or Markdown. Google Docs or Word: download as .docx.The files, or a ZIP of them

How Gurusup reads each export is in Exports from other tools.

Check that your files arrived

Open Ingest

Set the time range to cover your upload and pick Local Upload in the source filter.

Find your files

Each file, and each file inside an archive, is a row that reads "Ingesting", the file name, then "from Local Upload".

Watch the status

You should see it move from Queued to Learning to Learned. Every status is explained in Follow ingestion.

Members cannot open Ingest. They check Sources and its Local Uploads row.

On Sources, the Local Uploads row shows its status and its contribution, which grows as Gurusup reads. Local Uploads is created with your first accepted upload, with the access you chose then. It has nothing to reconnect or sync: its page shows a short description, Uploaded files with an Upload files button, and Knowledge access, where you can change who reads every uploaded file. Upload again to add more.

Replace or remove a file

Gurusup treats a file's path as its identity. For a file you upload directly, the path is its name. For a file inside an archive, the path starts with the archive name, for example handbook.zip/hr/leave.pdf.

You uploadWhat Gurusup does
The same path with the same contentRecognises it as unchanged and does not learn it again.
The same path with different contentTreats it as a replacement of that path.
Two entries with the same path in one selectionBlocks the upload and lists them as colliding.
A member uploads the same path as someone elseAdds it to the member's own upload source. It never replaces another person's file.
A later export where a page was renamedAdds a new document. The old one stays.

The app does not show which case applies before you confirm, or what a replacement does to knowledge learned from the earlier version.

You can remove a file only until you build your memory for the first time. After that, the app has no way to remove an uploaded file or what Gurusup learned from it, and Local Uploads has no Remove source action.

Before the first build, click Remove on the file's row in the Upload documents panel. In the Sources dialog, the X on a row only takes the file out of your selection before you confirm. A failed removal says "Could not remove" and the file name. After the first build, a removal is refused with "Files can only be removed before building your memory."

What you can upload

TypeNotes
Word, PowerPoint, Excel (.docx, .pptx, .xlsx)Gurusup reads the text. A workbook must have saved results for its formulas. Upload files and Send files on a webhook source both accept these.
PDFGurusup reads the text. A PDF with no text (for example a scan) is skipped. A PDF that is encrypted but opens without a password and allows text extraction is read like any other. A PDF that needs a password to open, or that does not allow text extraction, is refused.
Web pages (.html, .htm)Gurusup keeps headings, lists, tables and links. It drops scripts, styles, navigation menus, breadcrumbs and page footers, but keeps a footer inside the article. Pages saved in other encodings keep their characters. A page with no readable text is skipped.
Markdown (.md, .markdown) and text (.txt)Must be saved as UTF-8.
CSV (.csv)Must be saved as UTF-8.
JSON (.json)Must be valid JSON saved as UTF-8. A Roam or TiddlyWiki export is recognised and split into pages.
Notes from other tools (.org, .adoc, .canvas, .tid, .textpack)Org-mode and AsciiDoc files are kept as written. The others are read as in Exports from other tools.
Wiki and notebook exports (.xml from MediaWiki, .enex from Evernote)One document per page or note. An .xml file that is not a MediaWiki export is skipped.
Archives (.zip, .tar.gz, .tgz)Gurusup opens them and reads every supported file inside. A BookStack Portable ZIP is recognised too. See Archives.

Anything else, such as images, old .doc or .xls files, or audio, is skipped, and the dialog says "[file] is a kind of file Gurusup can't read. Save it as a PDF, Word or Excel file and add it again." A selection made only of unsupported files cannot be confirmed.

You cannot drop a folder: put it in a ZIP or TAR.GZ archive, and Gurusup keeps the folder structure in each file's path. There is no field for a URL: save the page as a PDF or Markdown file, or connect the app it lives in. For how file types compare with other sources, see Content types and limits.

Archives

  • Gurusup opens archives inside archives, down to four levels counting the outer one.
  • Each file inside is its own document, at a path that starts with the archive name.
  • Two entries whose paths match, ignoring capital letters, count as a collision and stop the upload.
  • An archive with a password-protected file, a symbolic link or another unsafe entry is refused whole.
  • A PDF inside an archive follows the PDF rules above, one file at a time.

Exports from other tools

Most exports give one document per page, note or card. Each one counts as an entry in the archive limits, and its size counts toward the total. Gurusup cleans generated names: slashes, backslashes and control characters are replaced, an empty name becomes "Untitled", and names that differ only by capital letters, such as Notes and notes, become Notes.md and notes (2).md.

ExportHow it comes inWhat Gurusup skips
BookStack Portable ZIP (book, chapter or page export)One document per page, at a path like Book/Chapter/Page.md (the chapter only when the page is in one), titled with the page name. Each starts with the name, the breadcrumb and the tags, then the content and the attachments. Link attachments keep their URL and file attachments are read by their type. Book and chapter descriptions become documents when they have text. Links between pages point to the new paths.Images, covers and files nothing refers to
MediaWiki XML (Special:Export)One document per page, at file.xml/Page title.wiki, titled with the page title. Each starts with the title and a "Last edited" date, then the wikitext as written, with templates left unexpanded. Only the latest revision is read. Talk pages stay.Redirects, wiki machinery (interface messages, templates, modules, translations and gadgets), content that is not wikitext such as Lua, JSON or CSS, and deleted or empty revisions
Evernote ENEXOne document per note, at file.enex/Note title.md. Each starts with the dates, tags, author and source URL. Checklists show as [x] and [ ]. Evernote 10 tasks are listed under Tasks with their status and due date.Notes with no text and no tasks. Attachments and images are named in the note, for example [attachment: wiring.png], and listed under Attachments, but never imported. Encrypted text shows as "[encrypted text, not imported]".
Roam JSONOne document per page, at file.json/Page title.md, with blocks as a nested list. Links and tags stay as written. A block reference, an embed or an alias becomes the quoted text followed by its page, for example “text” (from [[Page]]), up to 200 characters and one level deep. Daily notes are normal pages.Pages with no text
TiddlyWiki JSON, and .tid filesOne document per text tiddler. Its fields come first as key: value lines, then the body.System tiddlers (names that start with $:/), drafts, tiddlers that are images or other binaries, empty tiddlers, and the bag and revision fields. Single-file TiddlyWiki HTML is not supported.
Obsidian Canvas (.canvas)One section per card (text, file, link or group), then the connections with their labels.Nothing
TextBundle (.textpack, or .textbundle folders inside a ZIP, from Bear, Ulysses or iA Writer)The document is text.md, text.markdown or text.txt, titled with the bundle name.info.json and the assets. A bundle with no text file is skipped.
Logseq and Org (.org)An .org file uploaded alone is kept as written. Inside an archive, when a file has block IDs, a block reference or embed becomes the quoted text followed by its page, as for Roam, and the id:: and :id: lines are removed. A namespaced file name such as setting___enable journals.md becomes the page setting/enable journals, unless the file sets its own title::.The ID lines, inside an archive

Any other JSON file, and any recognised export that gives no pages, is read as plain JSON, so nothing is lost. A renamed page in a later export becomes a new document, see Replace or remove a file.

Limits

LimitValue
Size of one file you upload256 MiB (268,435,456 bytes). Checked in your browser first: a bigger file shows its reason and Choose another file on its row, and the rest of your selection stays.
Size of one document, after expanding64 MiB. This applies to a file you upload directly too, so a single PDF, CSV or Markdown file over 64 MiB is refused. Only an archive can be larger.
Total expanded size of one upload1 GiB
Files and folders inside one upload100,000 entries
Compression ratio200 to 1
Archives inside archives4 levels, counting the outer one
Depth of a path32 parts (folders plus the file name)
Length of a file path1,024 bytes. Inside an upload, a path of more than 500 characters is refused.
Files in one upload request1,000. The app sends your files one at a time, so you are unlikely to meet this.

A document Gurusup generates from an export, such as one wiki page, counts as one entry, counts toward the 1 GiB total and has to fit in 64 MiB.

Refusals and fixes

Messages that name a file show its path where this table shows <path>. In the Sources dialog and in the setup panel, the message sits in red under the file name, with the button described in How the app shows them.

Over a limit

MessageFix
"Upload exceeds the 256 MiB limit. Choose a smaller file."Split the file or compress it.
"This file is larger than the upload limit. Upload a smaller file."Same.
"These files are larger than the upload limit. Upload fewer or smaller files."Upload fewer files at a time.
"At most 1000 files may be uploaded."Upload fewer files at a time, or put them in an archive.
"A file exceeds the expanded size limit."A file, or a file inside an archive, is over 64 MiB when opened. Split it.
"This document contains more text than the upload limit. Split it into smaller files and upload them."A Canvas or TiddlyWiki file turns into more than 64 MiB of text. Split it.
"The uploaded files exceed the total expanded size limit."The upload is over 1 GiB when opened. Split it into several uploads.
"The uploaded archives exceed the entry count limit." or "The upload exceeds the file count limit."Split the upload into several archives.
"An archive entry exceeds the compression ratio limit." or "An archive exceeds the compression ratio limit."Re-create the archive without the highly compressible file.
"The upload exceeds the nested archive depth limit."Unpack the inner archives and upload their contents.
"The upload exceeds the file path depth limit."Flatten the folders.
"An uploaded file path exceeds the name length limit."Shorten folder and file names.
"A file path inside this upload is too long. Shorten its folder or file names and upload it again."Shorten folder and file names.
"Upload exceeds the 25 MiB (26,214,400 bytes) combined limit. Choose fewer or smaller files."Can appear when Gurusup uses its fallback upload route. Upload fewer or smaller files.

The file itself

MessageFix
<path> needs a password to open. Upload a copy that opens without a password.Save a copy of the PDF that opens without a password. A PDF that is encrypted but opens without one is accepted.
<path> does not allow text extraction. Ask its owner for an authorized copy that permits text extraction, then upload it.The PDF forbids copying its text. Ask whoever made it for a copy that allows it.
"More than one file in this archive is named" followed by the names, then "Rename them so each name is unique, then upload the archive again."Two files in the archive share a name. Rename one and create the archive again.
"Gurusup found no text to read in this file. Choose a document that contains text."Choose a document that has text in it.
"This archive has password-protected files. Remove the password and upload it again."Re-create the archive without passwords.
<path> can't be read as a PDF. Check that it opens, then upload it again.Open it yourself. If it opens, export it again as a new PDF.
<path> can't be opened. Open it in Word, Excel or PowerPoint, save a new copy and add that.Open and re-save it in Word, PowerPoint or Excel.
<path> has formulas without saved results. Open it in Excel, save it, and add it again.Open the workbook in Excel, let it recalculate and save.
<path> can't be read as text. Save it as plain text (UTF-8) or as a Word file and add it again.Re-save the file with UTF-8 encoding, or as a Word file.
<path> can't be read as JSON. Check its syntax and upload it again.Fix the JSON, for example with a validator.
<path> can't be opened as a ZIP archive. Re-create it and upload it again.Create the ZIP again. The same message exists for TAR.GZ.
<path> can't be read from its archive. Re-create the archive and upload it again.Create the archive again. The same message exists for TAR.GZ.
"The upload contains an unsafe file path."The archive holds an absolute path, a .. segment, a backslash or a drive letter. Re-create it with relative paths.
"The upload contains an unsafe archive symlink." or "... an unsafe archive link." or "... an unsafe special archive entry."Re-create the archive without links or special files.
"This file name can't be used. Rename the file and upload it again."Rename the file. A name cannot contain slashes or backslashes, or be only . or ...
"Gurusup cannot read this file. Check its format or choose another file."Shown on a row when Gurusup refuses the file and gives no other reason. Click Choose another file and pick a supported one.
"None of the selected files can be read by Gurusup. Choose PDF, Word, PowerPoint, Excel, Markdown, text, CSV or JSON files, or an archive of them."Add at least one supported file.

Who can upload, and what is off

MessageWhat to do
"Your organization access does not allow uploads. Ask an administrator to restore your access."Your membership is not active. Ask an admin.
"Uploads are paused for this organization. Ask an administrator when you can add files."Ingestion is paused. Wait until it resumes.
"Uploads are turned off in Demo Mode. Ask an administrator to turn off Demo Mode to add files."Demo Mode is on. Ask an admin.
"These access settings belong to another person. Choose who can read your files again."Close the dialog, open it again and choose who can read the files.
"Access to these uploads has changed. Choose different access settings or ask an administrator of your organization for help."Choose different access settings. If it keeps happening, ask an admin.
"Uploads with these access settings are no longer available. Ask an administrator of your organization for help."Ask an admin.

Skipped instead of refused

These do not stop the upload. Each one shows in the dialog as a line that names the file and says what to do, and the file is skipped. When the whole file cannot be read, the reason shows on its row instead.

ReasonFix
<path> is a kind of file Gurusup can't read. Save it as a PDF, Word or Excel file and add it again.The type is not in What you can upload, or an .xml file is not a MediaWiki export.
<path> has no text Gurusup can read, which usually means it's a scan. Export it with selectable text or save it as a Word file, and add it again.The PDF holds no text. Run OCR to make a searchable version, or save it as a Word file.
<path> has no text to read. Check that it has content, then add it again.A Word, PowerPoint or Excel file with no text in it.
"This HTML file has no readable text. Export a version with the page content, then upload it again."Export the page with its content.
"Single-file TiddlyWiki HTML isn't supported. Export your tiddlers as JSON from TiddlyWiki instead."Export from TiddlyWiki as JSON.
"This Canvas file can't be read. Export it again as JSON Canvas 1.0 and upload it."Export the canvas again in that format.
"This TiddlyWiki file can't be read as text. Export the note as Markdown or plain text and upload it."Export the note as Markdown or text.
"This TextBundle has no readable text file. Export it with text.md, text.markdown or text.txt and upload it again."Export the bundle again with a text file inside.
"This BookStack export can't be read. Export it again as Portable ZIP and upload it."Export again from BookStack as Portable ZIP.
"This MediaWiki export can't be read. Export the pages again from Special:Export and upload the XML file."Export the pages again and upload the XML file.
"This MediaWiki export has no pages with text. Export content pages, not only templates or redirects, and upload it again."Export content pages.
"This Evernote export can't be read. Export the notebook again from Evernote as ENEX and upload it."Export the notebook again as ENEX.
"This Evernote export has no notes with text. Export a notebook that has written notes and upload it again."Export a notebook that has written notes.

When it is on our side

MessageWhat to do
"Uploads aren't available right now. This is on our side, and nothing was uploaded. Contact support and we'll sort it out."Contact support.
"Gurusup can't read PDF files right now. This is on our side. Contact support and we'll sort it out." (same for Office files)Contact support.
"Gurusup couldn't store this file. This is on our side. Try again in a few minutes, and contact support if it keeps failing."Try again later.
"Gurusup couldn't save this preview. This is on our side, and nothing was uploaded. Try again in a few minutes, and contact support if it keeps failing."Try again later.
"Your files are safely stored, but Gurusup couldn't start reading them."Confirm the same selection again. Nothing is uploaded twice.
"Uploads are paused while we update your organization. Try again in a few minutes."Try again in a few minutes.
"This file didn't finish uploading. Choose it again to restart the upload."Add the file again.
"This preview has expired. Preview the file again." or "These files changed since you previewed them. Preview them again."Close the dialog and add the files again.

How the app shows them

Each file fails on its own row, and the message is the reason itself, with no prefix. The button under it depends on the cause.

CauseButtonMessage
The connection or the upload service is temporarily unavailable (no connection, or a 408, 429, 502, 503 or 504 answer)Retry"Could not connect to upload the file.", or the reason Gurusup gives. With none: "The file could not be uploaded because the connection or upload service is temporarily unavailable. Try again when the connection is available."
Gurusup refused the file (a 400, 413, 415 or 422 answer), or your browser did: the file is over 256 MiBChoose another fileThe reason from Refusals and fixes, for example "Upload exceeds the 256 MiB limit. Choose a smaller file." With none: "Gurusup cannot read this file. Check its format or choose another file."
Anything else, such as 403 when uploads are turned offNone. Use RemoveThe reason Gurusup gives. With none: "Gurusup could not prepare this file. Your selection is still here. Contact support for help." While confirming it says "confirm" instead of "prepare".

Choose another file opens your file picker, and the file you pick takes the place of that row. In the Sources dialog, an error while you confirm shows as a message that starts with the file name, and your selection stays.