Minds shows most problems as a short notice, a badge on a chat or a hover text on a control you cannot use. The AI assistant panel that runs next to other screens works the same way. This page lists the messages you may see, what each one means and what to do about it.
| Permission | What it lets you do |
|---|---|
| Add Chat/Message | Send messages in your own chats. |
| Add Message In Others Chat | Send messages in a chat that someone else owns. It covers sending only, not renaming or deleting. |
| Create Private Chat | Start a private chat, which only you can see. |
| View Training Cards | Convert a shared training card to a local card. |
See Who can do what with Minds for the full list.
| You see | What it means and what to do |
|---|---|
| No Minds Available | No Mind is open to you, or none exists yet. Create your first Mind if you may; otherwise ask to be added to a Mind's owner role, or ask your administrator to create one for you. |
| Access denied to this mind | You tried to start a chat with a Mind you cannot reach. Ask to be added to its owner role. |
| Forbidden with You do not have access to this conversation. | You cannot see this chat. Private chats are visible only to their creator. Ask the owner. |
| You do not have permission to create private chats. | Ask for Create Private Chat, or switch the chat to Regular. Copying a shared private chat needs the permission too. |
| You don't have permission to send messages here. Ask your administrator for the "Add Chat/Message" permission. | The message box is greyed out. Hover over it to read the reason, then ask for the named permission. The text names Add Message In Others Chat in someone else's chat. |
| You see | What it means and what to do |
|---|---|
| Loading chat | Minds is opening a chat and has not finished loading the Mind and its conversation. Wait; it clears on its own when the chat is ready. |
| Show panel | The top-left button of Minds on a phone. It opens the left panel with New Chat, Search, Minds, Artifacts and Scheduled Tasks, then your chats. Tap it; picking an item closes the panel again. |
| Show Tool History / Hide Tool History / Show Artifacts / Hide Artifacts | The chat header buttons that open or close the right panel (Ctrl+Shift+H and Ctrl+Shift+F). Opening one panel closes the other. |
| Parent conversation | In a swarm agent's chat, the parent conversation's name is not loaded yet; the label normally reads Parent: {name}. Click it to open the parent; the name appears once the swarm tree loads. |
| No minds found | The Minds list or the search found no Minds: a search, the bookmarks toggle or the filters hide everything, or there are none to show. Clear the search and filters, switch the bookmarks toggle off, or create a Mind if you may. |
| No matching minds found | The search in the compact Minds list matched no Mind name. Change the words, or clear the search with Clear search. |
| No bookmarked minds found | The bookmark filter is on and the list holds no bookmarked Minds. Switch the filter to Show all minds, or bookmark a Mind. |
| No agent minds found | The compact list has no Minds to show, with no search and no bookmark filter on. Create a Mind from Add Mind if you may, or ask your administrator. |
| Loading mind | Placeholder rows shown while the Minds list or a Mind form loads. Wait for the list to finish loading. |
| Search chats, minds, artifacts… | The placeholder of the search dialog's input; it opens with Ctrl+Shift+S. Type to search chats, Minds and artifacts; clear the field to return to the quick actions and recent chats. |
| Untitled Conversation | A chat in the Recent Chats list has no title. Nothing to fix; open it from the list. |
| No results found / Try different keywords | The search dialog found nothing for the words you typed. Clear the search and try different keywords; a recent chat can still be opened from the list. |
| No chats found | No chat matches your words. Try different keywords. |
| No artifacts found | No artifact matches your words. Try different keywords, or open the Artifacts area. |
| No conversations yet / Start a new conversation | The list is empty; nothing is broken. Start a chat with New Chat (Ctrl+Shift+Q) and it appears here. |
| Failed to load conversations with Please try again | The chat list could not load. Try again; if it persists, reload the page. |
| This action permanently removes {title}. You cannot undo this. | The delete confirmation for a chat; deleting a chat is permanent. Confirm only when you no longer need it; nothing can bring it back. |
| Scroll down | There is newer content below and the chat stopped following as you read; the button shows only while a response streams and you have scrolled away from the bottom. Click it to jump back down; the chat follows the answer again. |
| View swarm agents | Opens the swarm agents tree; the button appears only when the chat has swarm agents, and only on a desktop screen. Click it, or open the tree from the swarm bar above the message box. |
| View message hierarchy | Opens the Message Hierarchy map; it is hidden on a phone. The map's legend reads Active path and Inactive. |
| Jump to branchpoint / Jump to swarm report / Jump to message | What a minimap dot does when you hover it. Click the dot; a target that is not loaded yet is loaded as you jump. |
| Click any message to navigate | The Message Hierarchy map's hint. A click on the active path closes the map and scrolls to that message; a click elsewhere switches the chat to that branch. |
| Active path / Inactive | The map's legend: the path you are reading versus the other paths. Click a message on another path to switch to it. |
| No messages yet / Start a conversation to visualize the message hierarchy | The Message Hierarchy map has nothing to show. Send a message in the chat and open the map again. |
| You see | What it means and what to do |
|---|---|
| Message not sent | The platform refused your message. Read the text under the title, fix the cause and send again. |
| Failed with There was an error generating response | The answer failed. Click Retry to run the turn again, or Continue (shown when part of the answer arrived) to finish it. |
| Retry not started / Continue not started | The retry could not start. Read the description and try again. |
| The server restarted while this response was being generated. Press Retry to run the turn again or Continue to resume it. | A platform restart stopped the answer. Click Retry to run the turn again, or Continue to resume it. |
| Connection lost with Trying to reconnect. Sending is paused until the connection is back; reload the page if this does not clear. | The live connection dropped and sending is paused. Wait for it to return. If it does not clear, reload the page. |
| Not connected with The chat connection is down. Your edit is kept - try again once it is back. | You edited a message while offline. Your edit is kept. Save it again when the connection is back. |
| name is currently processing a request with Please wait for them to finish (fallback A user is currently processing a request) | Another person is sending in this chat; one turn runs at a time. Wait for their answer to finish. |
| Another user is currently writing | The same as above, shown when a send is refused. Wait, then send again. |
| Model locked for this conversation | The chat is locked to one model, so your choice was overridden and the picker snaps back to the locked model. Keep the locked model or start a new chat. |
| Your message is back in the input. / Your attachment is back in the input. / Your attachments are back in the input. / Your message and attachments are back in the input. | Restore on a failed-send notice put the unsent message or attachments back in the message box. Edit it and send again. |
| Conversation not found | This chat no longer exists, so the action did not run. Open another chat from the conversation list. |
| Message not found | This message no longer exists, so the action did not run. Reload the chat and try again. |
| You don't have permission to send messages in this conversation. | You cannot send in this chat, so the change was not saved. Ask your administrator for the send permission. See Who can do what with Minds. |
| Limit Reached with You can only attach up to 2 {type} artifacts. / Maximum 2 allowed per message | A message can hold two artifacts of one type; the card counts the artifacts of that type already attached. Remove one before you attach another, or send it in a second message. |
| Already Attached with This {type} configuration is already attached to your message. | The artifact is already in this message. |
| Cannot cancel with This question is missing a tool id. Try refreshing the chat. / Could not cancel question | The Mind's question could not be skipped. Read the description. If it says to refresh, refresh the chat and try again. |
| Could not change mode | The change to acceptance mode failed. Read the description and try again. |
| Detecting PII... | Your message is being checked for personal data before it reaches the model; the label replaces the usual thinking phrases. Wait; nothing is needed from you. |
| Waiting for Tools | The Mind is running tool steps for this turn; up to five tool icons show what is running, then a counter for the rest. Wait for the steps to finish. |
| Generating Response | The answer text is being written. Wait; the answer appears as it streams. |
| Compacting conversation… | The chat is summarising its older turns to free up room. Wait; the pill clears when it finishes and a summary card appears. |
| Conversation compacted with n messages summarized | A summary replaced older turns of the chat. Expand the card to read the summary. |
| Context cleared with The next reply starts fresh. | The context was cleared, so earlier messages are hidden from the assistant while staying visible to you. Keep chatting; the next reply starts from a blank context. |
| Delivered artifacts | Swarm agents finished work and handed results back into this chat. Read the delivered results in the parent chat; its artifact lists are updated. |
| Private with Only you can see this conversation in shared lists. Messages stay between you and this mind. | The chat is private. Nothing required; switch the chat to Regular if others should be able to help with errors. |
| Loading earlier messages | Older turns are loading as you scroll up. Wait a moment. |
| No feedback yet | Nobody has rated this answer yet; the feedback list attached to the message opens empty. Rate the answer with the like or dislike control, or wait for other people's feedback to arrive. |
| Switch to an available model to edit | The chat's model cannot be used right now (it is retired, not available, or cannot read the files in the chat), so editing is off; this is the tooltip on the greyed Edit message control. Switch the chat to an available model and try again; if the chat is locked to its model, start a new chat. |
| Stop the current response with Stopping... | The send button has become the stop button while your answer streams. Click it to stop the answer; it shows Stopping... until the answer stops. |
| The message is too large to send. Shorten it or remove some attached text, then try again. | The message would exceed the size the chat can send in one piece, so nothing was sent. Shorten the text or remove some attached text, then send again. |
| Plan · {done}/{total} with Completed / In progress / Pending / Skipped | The Mind's plan for the current work and the state of each step. Open the bar to see each step; a finished plan folds to all steps done and clears after your next message. |
| You see | What it means and what to do |
|---|---|
| Nothing to compact with Open a conversation with some history first. | / needs an open chat. Open one, or use it after a few turns. |
| Nothing to compact yet: the conversation is still short. | The chat is too short to summarise. Keep chatting. |
| Wait for the current response to finish before compacting. / Wait for the current response to finish before clearing the context. | The command was used during an answer. Wait for the answer, then try again. |
| Compaction failed / Context not cleared | The command failed. Read the description and try again. |
| Context not cleared with Not available in a swarm agent chat. | / does not run in a swarm agent chat. Use the parent chat. |
| Context not cleared with Answer the pending question first. | Answer the Mind's question, then run the command again. |
| Context not cleared with The chat connection is down. Wait for it to reconnect and try again. | Wait for the connection to return, then run the command again. |
| Nothing to clear with Open a conversation with some history first. | / needs a chat with history. Open one first. |
| No matching command. Enter sends the text as typed. | The letters after / match no command. Fix the spelling, or press Enter to send the text as an ordinary message. |
| This mind has no tables to reference. | The Mind has no tables the @ picker can show. Link tables on the Mind's Sources tab and keep Show in chat @ picker on; tables that are not deployed or have no readable columns stay out. |
| Could not load the mind's tables. | The picker could not load the Mind's table list. Try again in a moment; if it keeps failing, check the Mind's sources. |
| You see | What it means and what to do |
|---|---|
| Blocked: personal information detected with Your message was not sent. It looks like it contains {what}. Edit it to remove or mask that, then send again. | PII enforcement in block mode refused the send; the sentence names the kinds of personal data that were found (for example "an email address"). Edit or mask the highlighted parts and send again; a blocked message cannot be sent as it is. |
| Warning: personal information detected with Your message was not sent. It looks like it contains {what}. Edit it to remove or mask that, then send again. | Warn mode held the message back, but the exact same text may be resent with the bypass by a permitted user. Edit the text, or, in warn mode and with the Bypass PII Detection permission, choose Send it anyway; any edit disarms the bypass. |
| Personal information detected with This message was not sent. Edit the highlighted parts to remove or mask them. | The details dialog for a flagged message, opened by tapping the notice; it shows the flagged parts highlighted. Use Edit message to return to the message box, Close to dismiss, or (warn mode and permission only) Send it anyway. See PII checks in chat. |
| Send it anyway with This message contains flagged personal data. Sending now skips the check - or edit the highlighted parts first. | The send button is armed to resend the exact warned text past the check (warn mode and the bypass permission; the screen-reader label reads "Send anyway (skips the PII check)"). Press it to send unchanged, or edit the highlighted parts first, which cancels the bypass. |
| You do not have permission to send it this way. Edit the message and send again. | A bypass send was refused because the account lacks the PII bypass permission. Edit the message to remove or mask the personal data and send again, or ask an administrator for the bypass permission. |
| No PII detected with Your draft looks clean against the platform policy. | The on-demand draft check found no personal data. Send the message as usual. |
| PII scan failed | The draft check could not run; the notice carries the reason the server gave when there is one. Run the check again; if it keeps failing, tell your administrator and quote the message. |
| No PII detected with Nothing in this text matched the enabled detection rules. | The full PII scan dialog found nothing in the checked text. Nothing to do. |
| Scanning… | The full PII scan dialog is running the check; it starts as soon as the dialog opens. Wait for the result. |
| You see | What it means and what to do |
|---|---|
| You cannot share a conversation with yourself. | You picked yourself as the person to share the chat with. Choose another person; the chat is already yours. |
| A share may last at most 90 days. | The end date is more than 90 days away (a date that is not in the future is refused too). Pick a date within 90 days from today. |
| This conversation was shared with you read-only. You cannot send messages in it. | The viewer is on a share, and a share never grants writing. Read the chat, or start your own copy of it. |
| A conversation shared with you cannot be shared onward. | The sharer reached the chat through a share themselves; a share is one hop. Ask the person who shared it with them to share the chat with the other person. |
| Only a chat that was shared with you can be copied. | The copy action is for share guests only; the owner and users who read the chat through a role cannot copy it this way. Nothing to fix on your side; the owner carries on in the original chat. |
| You cannot start chats on this mind. | The copy cannot be created because you cannot start chats on that Mind. Ask for access to the Mind that allows starting chats. |
| Only the main chat can be copied, not a sub-chat. | The shared chat is a sub-chat, and only the main chat can be copied. Copy the main chat instead. |
| Sub-conversations cannot be shared directly. | You tried to share a sub-chat. Share the main chat instead. |
| This chat was copied from a shared chat. Spreadsheets from the original chat are not loaded here, so I can't run new queries on them. To ask new questions about a spreadsheet, upload it again. | A copied chat cannot query the spreadsheets that were uploaded in the original chat. Upload the spreadsheet again in the copy. |
| You see | What it means and what to do |
|---|---|
| Model retired / "{name}" is retired and no longer available. Switch to another model to continue chatting. | The model is no longer available. Switch to another model to continue; if the chat is locked to it, start a new chat. |
| Model unavailable / "{name}" isn't activated for this mind. / "{name}" isn't activated for this mind and this conversation is locked to it. Activate it in the mind's AI Config, or start a new chat. | The model is not turned on for this Mind. A Mind manager can activate it in AI Config; people who cannot edit the Mind see "Start a new chat, or ask a mind manager to activate it." If the chat is locked to it, start a new chat. |
| Model incompatible / This conversation contains {images|PDFs|images and PDFs} that "{name}" can't read. Switch to a model that supports {images|PDFs|images and PDFs} to continue chatting. | The chat holds images or PDFs that the selected model cannot read. Switch to a model that reads them, then send. |
| This conversation is locked to "model" / This conversation can't switch to a Kartlos model / This conversation contains {images|PDFs|images and PDFs} this model can't read | The model picker row is disabled for this chat: it is locked to another model, cannot switch to a Kartlos model, or cannot read the files the chat holds. Start a new chat to use another model. |
| Model(s) disabled by your administrator: names. Ask an admin to allow them, or pick from the available model list. | A model you chose is not allowed for this company. Remove it, or ask an administrator to allow it. |
| Unknown model(s): names. Pick from the live model catalog. | The model is no longer in the catalog. Remove it and pick from the list. |
| A message that starts Model catalog is unavailable | The list of models cannot be loaded right now, so the model selection cannot be checked. Try again shortly. |
| At least one model must be available / Default model is required / Max concurrent agents is required | A required field in AI Config is empty. Fill it in. |
| Select model | No model is set for the chat yet. Pick one from the list; the first message you send fixes the chat's model. |
| Not available (on the Effort control) | The selected model has no thinking-effort options, so there is nothing to set. Nothing to fix if you do not need thinking effort; switch to a model that supports it if you do. |
| No models available to compare. | The AI Models Comparison window found no models to list. Close the window and use the chat's model picker as usual. |
| You see | What it means and what to do |
|---|---|
| Attachments unavailable with The selected model can't read images or PDFs. Switch to a model that supports them to attach files. | The Attach control is locked because the selected model reads no images or PDFs and Excel or CSV uploads are off for the Mind. Switch to a model that supports them, then attach the file again. |
| Unsupported file with File "{name}" is not supported. | The file type is not allowed. Use a JPEG, PNG, GIF or WebP image, a PDF, or an Excel (.xlsx or .xlsm), CSV or TSV file. |
| Unsupported file with Legacy .xls is not supported. | Save the file as .xlsx or .csv, then upload it again. |
| Not supported by this model with "{name}" can't be read by the selected model. Switch to a model that supports images or PDFs to attach it. | The selected model cannot read this kind of file; the check runs before the size and count checks, so a small file can still be refused. Switch model, then attach the file again. |
| Attachment limit reached / Maximum 20 images per message. Please remove some images first. / Maximum 5 PDFs per message. Please remove some PDFs first. / Maximum 5 Excel / CSV files per conversation. Please remove some Excel / CSV files first. | A message holds up to 20 images and 5 PDFs; a conversation holds up to 5 Excel or CSV files. Remove some first, or start a new chat for the next spreadsheet. |
| Image too large / Image "{name}" is too large. Maximum size is 3.75 MB per image. / Image "{name}" dimensions ({w}x{h}px) exceed the maximum allowed size of 8000x8000px. | An image must be 3.75 MB or smaller, and no more than 8000x8000px. Shrink, compress or crop it below 8000 px per side. |
| PDF too large / PDF "{name}" is too large. Maximum size is 30 MB per PDF. | A PDF must be 30 MB or smaller. Split or compress it. |
| PDF exceeds maximum page limit. Maximum allowed: 100 pages, uploaded: {pageCount} pages | The PDF is longer than 100 pages. Split it and attach the parts you need. |
| Excel / CSV file too large / File "{name}" is too large. Maximum size is 15 MB per Excel / CSV file. | An Excel or CSV file must be 15 MB or smaller. Split it, then attach the part you need. |
| Excel / CSV uploads disabled / Excel / CSV uploads are disabled for this Mind | Excel and CSV uploads are off for this Mind. Ask a Mind manager with Edit Mind to turn on Allow Excel / CSV uploads, or load the file into the Lake. |
| Excel / CSV attachments don't transfer between Minds | Switching to another Mind removed the Excel and CSV attachments from the draft. Attach the file again in the new chat; images and PDFs move with you. |
| Preparing tables… | An attached Excel or CSV file is still being turned into tables. Wait; sending stays blocked until the file is ready. |
| Ready · {n} table(s) | The file is ready and became this many queryable tables. Send your question. |
| Failed · {error} | The file could not be prepared; the specific error follows the dot. Remove the file and upload a corrected one. |
| Could not prepare tables | The file could not be turned into tables. Remove it and upload it again, or save it as .xlsx or .csv and retry. |
| Excel / CSV attachment not ready with Review PII columns before sending | Review the personal-data columns in the dialog, then send. |
| Excel / CSV attachment not ready with Remove the failed Excel / CSV attachment before sending | Remove the file that failed, then send. |
| Excel / CSV attachment not ready with Excel / CSV attachments are still being prepared | Wait until the file is ready, then send. |
| Only the uploader can confirm PII columns for this attachment. | The personal data review belongs to the person who attached the file. Ask the uploader to finish the review. |
| Detection here is regex-based only (no ML). Mark every column that holds personal data; it will be tokenized before the model sees it. | The review's suggestions are pattern-based, so the results must be checked by a person. Mark every personal data column (or declare the file clean), then confirm. |
| Bypass is recorded in the PII detection log. | Using the bypass and sending raw is logged. Use the bypass only when the file is safe to share with the model. |
| Could not save PII columns / Could not bypass PII review | The review did not save. Try again, or mark the columns instead. |
| Upload Failed / Unable to upload | The upload did not finish. Try again, or upload again from the same chat. |
| Loading preview... | The preview is opening. Wait. |
| Preview not available | The file cannot be shown in the preview. Download the file instead. |
| Preview Failed with Failed to load preview. Please try again. | The attachment preview did not load. Close it and open it again, or download the file. |
| File has no header row | The sheet's first row holds no column names, so the file could not be read as a table. Put column names in the first row, upload the file again. |
| Too many sheets (max 20) | The workbook holds more sheets than one upload allows. Split the workbook and upload it in parts. |
| Microphone access denied with Enable in browser settings | Dictation cannot reach the microphone. Allow microphone access for this site in your browser settings. |
| Microphone access required with Microphone access is required for speech recognition. Please enable it in your browser settings. | The same fix: allow microphone access for this site, then start dictation again. |
| Voice input isn't available on iPhone or iPad. Try Chrome, Edge, or Safari on desktop or Android. | Dictation cannot run on that device. Use a desktop browser or an Android device. |
| Voice input isn't supported in Firefox. Try Chrome, Edge, or Safari. | The browser lacks the speech-recognition support the microphone needs. Switch to a supported browser. |
| You see | What it means and what to do |
|---|---|
| Not found | The reference could not be matched to an object or person, because it was deleted or it is not available to your account. Open the object from its own area (Warehouse, Lake, DataPrism or Workflows) if you still have access; the chat cannot open it. |
| User #{id} | The reference holds a user id that has no match, so no name can be shown. The name appears when the id matches a person. |
| Role #{id} | The role was deactivated or deleted, so only its id remains. A role chip has no target surface. |
| Workflow #{id} | The workflow's name has not loaded yet, or it could not be loaded. Wait for the name to arrive; open the workflow from Workflows if it stays an id. |
| Run #{id} | A workflow run has no name of its own, so the chip always shows its id. The chip opens the run. |
| ⚠️ Invalid artifact format | A reference inside an answer cannot be read as an artifact reference, so the raw text is shown in its place, or nothing when there is no text. Nothing is broken in the chat; the item can still be found under Artifacts. |
| You see | What it means and what to do |
|---|---|
| Values with ~ are estimated while the reply streams. | The figures shown while a reply is being written include a live estimate for the running turn; the settled figures replace the estimate once the reply finishes. Read the marked values as provisional and check the figures again after the reply finishes. |
| Swarm agents run as separate chats, not counted above. | Tokens spent by swarm agents are not included in the conversation's figures; the dialog lists them in a separate Agents row. Treat the Agents row as separate from the conversation's totals. |
| You see | What it means and what to do |
|---|---|
| The mind with the name "name" already exists | Choose a unique name. |
| Name must be at least 2 characters / Name must not exceed 60 characters / Owner Role is required / Responsible person is required | Fix the highlighted field. |
| Mind name is required | The Mind Name in the Settings form is empty, so the save does not go through. Enter a name, then save again. |
| Mind name must be less than 50 characters | The name failed the Settings form's length check (the field itself allows up to 60 characters, so the message reads stricter than the enforced limit). Shorten the name and save again. |
| Name cannot contain backticks, double quotes, colons or backslashes | The name uses a character that is not allowed. Remove those characters from the Name and create the Mind again. |
| Mind Color must be at least 2 characters | The form's check on the colour does not pass, so the update does not go through. Open Appearance and pick a colour, then click Update again. |
| Mind Icon must be at least 2 characters | The form's check on the icon does not pass, so the update does not go through. Open Appearance and pick an icon, then click Update again. |
| Cannot assign owner role "role": you do not hold it | You can only give a Mind roles you hold. Pick one of yours or ask an administrator; only admins assign the admin role. |
| Cannot assign permission group(s) you do not hold: roles | Remove the roles you do not hold. |
| A Mind can have either a responsible user or a responsible role, not both | Pick one. |
| Failed to create mind / Failed to update mind | The save did not finish. Try again; if it keeps failing, ask your administrator. |
| No data sources attached to this mind. Please add sources to enable chat functionality. | The Mind has no tables linked yet. Click Add Sources to link tables. |
| You do not have permission to manage this mind / Not authorized to update this AI mind / Not authorized to delete this AI mind / Not authorized to duplicate this AI mind / Not authorized to see this AI mind | You lack the permission for this action. Ask for the matching permission or role. For a Warehouse Builder Mind, only an admin can reach it. |
| AI mind with ID id not found / Mind "name" not found or not accessible | The Mind no longer exists or is not open to you. Refresh the list and pick another Mind; if it should be there, ask its owner or an administrator. |
| Source mind is missing. Close this dialog and try again. | The duplicate dialog lost its Mind. Close it and start again. |
| A Lake Builder Mind can only be duplicated as a Lake Builder Mind / A Warehouse Builder Mind can only be duplicated as a Warehouse Builder Mind | Keep the same kind when you duplicate it, or create a new Mind of the kind you want. |
| The selected owner role differs from the original mind's owner role ({name}). You may not have permission to access some of the data sources and information configured in this mind. | The duplicate's owner role is not the original's, so parts of the original's configuration may be out of reach. Keep the original owner role if unsure, or check with an administrator who can see the copy's data. |
| Mind was duplicated but the response did not include the new mind. Refresh the minds list. | The copy may exist even though the form did not return it. Refresh the Minds list and check for the copy before trying again. |
| Owned by the admin role · runs as the messaging user · scheduling stays manual / Lake Builder Minds are owned by the admin role. | The note a builder kind shows in the create form, the duplicate form and the Mind's settings, in place of the Owner Role picker. Nothing to set; the ownership, run-as and scheduling rules are fixed for the kind. Ask a person with the admin role to manage the Mind. |
| Lake Builder Minds always run as the messaging user. / Warehouse Builder Minds always run as the messaging user. | Query Identity is locked to Execute as User on a builder Mind; it always runs as the person messaging it. Nothing to change. |
| Only Lake Administrators can manage Lake Builder Minds. | Managing a Lake Builder Mind was refused without Add/Modify Lake Objects. Ask an administrator, or for the Add/Modify Lake Objects right. |
| Inherited from owner role | The chip is locked because the Mind uses the Owner Role query identity. The owner role's access is inherited. |
| No groups selected | The Mind has no permission groups yet. Click Add permission group to add roles. |
| Loading settings | The Settings tab is loading the Mind's values. Wait for the fields to appear. |
| Turn on Global PII detection first | Ask an administrator to turn on PII detection, which is needed for Enforce PII column clarification. |
| No executable role is available for this mind. | The Mind runs queries as you, and you hold none of the roles it can use. Ask for a role in its owner role or permission groups; on a builder Mind, check with your administrator if it persists. |
| SQL execution is disabled while impersonating another user; exit the view to run queries | You are viewing as another user. Exit the view to run queries; a Mind set to Owner Role keeps answering during the view. See Impersonate a user. |
| Delete Mind? with Do you want to delete this mind? | The confirmation before a Mind is deleted, opened from the list or the Mind panel. Click Delete to confirm; the Mind's chats, scheduled tasks, training cards, uploaded files and its table access are removed, and the Lake and Warehouse tables themselves stay. |
| Cancel mind setup? with This will delete this mind and cannot be undone. You can create a new mind anytime. | You clicked Cancel setup in the wizard, which deletes the new Mind. Choose Delete mind, Keep editing, or Finish Setup to keep the Mind. |
| Unsaved Changes with You have unsaved changes in the current step. If you navigate away, your changes will be lost. Are you sure you want to continue? | You tried to leave a wizard step with unsaved edits. Choose Discard changes, or Stay on this step to keep editing. |
| Discard unsaved changes? | You are switching tabs with unsaved edits; the dialog names the tab you are on and the one you asked for, and adds Leaving this tab will discard them. when you leave instead. Choose Discard changes to drop the edits, or Stay on this tab to keep them. |
| You see | What it means and what to do |
|---|---|
| Selection limit reached with Cannot select more than 100 tables | A Mind can have up to 100 tables. Deselect some. |
| Maximum 100 tables can be selected at once | The same limit, shown when you pick tables. Deselect some. |
| Failed to save sources | The sources were not saved. Read the description under the title and fix that cause (a missing column selection or a column the owner role cannot see). |
| At least one column must be granted to link a table to a mind | Select at least one column. |
| Table "name" must be deployed before it can be linked to a mind | Deploy the table first. See Configure a table or Deploy, run, truncate and monitor tables. |
| Columns not visible to the mind owner role cannot be granted: columns | Remove those columns from the selection. |
| Not deployed / This table isn't deployed, so this mind can't read it. | The table is no longer deployed, so this Mind cannot read it. Redeploy the table, or remove it from the Mind. |
| Training completed with Mind sources have been successfully retrained. | The Retrain run finished; the Mind has relearned its tables. Nothing to do. |
| Training failed | Retrain sources did not finish. Read the description and try again. |
| No Mind Sources Available with There are currently no mind sources available for this mind. | No tables can be linked. Check your data sources, or ask your administrator. |
| This Mind sees every Lake table; no sources need to be linked. / This Mind sees every Lake and Warehouse table; no sources need to be linked. | A builder Mind reaches every table on its own; sources are not linked to builder Minds. Nothing to do. |
| No Warehouse tables available / No Lake tables available | The edit view has no tables to show for that side. Check that deployed tables exist and the owner role can reach them. |
| No Warehouse tables assigned / No Lake tables assigned | The view mode has no linked tables for that side. Open Edit Sources and link tables. |
| The columns you select do more than train the AI. The AI SQL user is granted access to exactly these columns and nothing else. | The column ticks set the Mind's query access, not only its training data. Select only the columns the Mind needs. |
| Unsaved changes - the estimate refreshes after you save. | The injection-estimate banner does not reflect edits until you save. Save to refresh the estimate. |
| You do not have permission to view type tables / Table ID not found | You cannot open that table. Ask for view rights to its type. |
| You see | What it means and what to do |
|---|---|
| Training card not saved with the rejection reason; action Save it anyway | PII detection stopped the save; a warning-coloured toast can still be forced, a destructive-coloured one cannot. Edit the highlighted parts and save again, or, in warn mode and with the Bypass PII Detection permission, choose Save it anyway for the unchanged text. |
| Read-only with Shared-in cards can only be edited from their owner mind. | Edit the card in its owner Mind, or choose Convert to local. |
| Vectorization in progress with This card is being vectorized - try again in a moment. / Wait for the current vectorization to finish before saving. | The card is being processed for search. Wait, then try again. |
| PDF conversion in progress with Wait for the PDF conversion to finish before saving. | The PDF is still converting. Wait, then try again. |
| Nothing to save with No changes compared to the last saved version. | The card has no changes. Make the edit you intended, then save. |
| Cannot enable card with This card has no content to enable. | The card has no content, its PDF is still converting, or its PDF conversion failed. Add content, wait, or fix or remove the PDF. |
| Nothing to enable with Selected cards are still converting, failed, or have no content. | The selected cards are still converting, failed or have no content. Fix them first. |
| A training card with name "name" already exists in this mind | Rename the card. |
| Training card not found | The card may be deleted or still loading. Search for it again. |
| Upload blocked with An upload is already running | Wait for the current upload to finish. |
| Upload blocked with Only PDF, Markdown (.md) and text (.txt) files are allowed | Use a file of one of these types. |
| Upload blocked with PDF must be 30 MB or smaller / Markdown and text files must be 5 MB or smaller | Shrink or split the file. |
| Upload blocked with At most 50 files per upload | Upload fewer files at a time. |
| Uploading {n} files in the background with Keep this browser tab open until the upload finishes; you can keep working in DataMind. | The files upload one at a time in the background. Keep this browser tab open until the upload finishes. |
| {n} file(s) uploaded | Every file in the upload queue finished. Review the new cards in the Training Hub. |
| n uploaded, m failed | Some files did not upload. The notice lists the failed file names and the first failure reason; upload those again. |
| Upload batch finished with {n} ready, {m} failed. / {n} card(s) ready. | The batch's conversions finished. Review the new cards and enable the ones the Mind should use. |
| PDF converted with {card name} markdown is ready. | A card's PDF conversion finished and its text is in the card. Open the card and review the text. |
| PDF conversion failed with The PDF could not be converted to markdown. | Only text-based PDFs are supported. Upload another file, or upload this one again. |
| Conversion failed with Try another file or upload again. Only text-based PDFs are supported. / Try a different PDF | This card's conversion failed. Replace the file, or remove it and upload another one. |
| Conversion cancelled with PDF conversion was cancelled. | You stopped one card's conversion. The card stays; upload another file for it when ready. |
| Cancel remaining PDF conversions? with This stops conversion and deletes cards that have not finished. Already converted cards stay. | The rest of the batch will be stopped and unfinished cards removed. Confirm with Cancel conversions, or close the dialog to keep it running. |
| Batch conversion cancelled / Batch conversion partially cancelled | The batch stop went through for all or part of the batch. Already converted cards stay. |
| Invalid file / Upload failed / Remove failed / Cancel failed | Try again; for an invalid file, use a PDF, Markdown or text file. A batch cancel that fails reads Batch conversion could not be cancelled. |
| File uploaded with PDF uploaded and queued for conversion. | A single-file attach on a card was accepted and queued. Wait for the conversion to finish. |
| File content applied with The file text replaced this card content. | A Markdown or text file replaced the card's text. Review the card. |
| Attachment removed with The PDF was detached from this training card. | The file was removed from the card. Upload a new file if needed. |
| Wait for PDF conversion to finish before enabling. / Conversion failed - fix or remove the PDF first. / This card has no content to enable. | The card cannot be enabled yet. Wait for the conversion, fix or remove the failed PDF, or add content. |
| Preview failed / Thumbnail failed / Download failed | Try again. |
| Vectorization failed with The card could not be vectorized. / Vectorization failed - save the card again to retry. | Save the card again to retry; check the content if it keeps failing. |
| Vectorization complete with "{cardName}" was vectorized successfully. | The card is ready for similarity search. Nothing to do. |
| Delete failed / Partially deleted | Some cards were not deleted. Try again for the remaining cards. |
| Delete selected training cards? with This will permanently delete {n} selected card(s). Cards shared with other minds move to one of those minds instead. This action cannot be undone. | A batch delete is final; cards shared with other Minds are handed over instead of deleted. Confirm only when sure; remove the shares first if the cards should really be gone. |
| Delete folder? with This will remove the folder "{name}" and delete all training cards inside it, including nested folders. | Deleting a folder deletes the cards inside it, nested folders included. Move the cards you need out of the folder first. |
| Share failed / Unshare failed / Copy failed / Convert failed | Read the description and try again. The copy failure details name conflicts in target Minds, or a card still converting or being vectorized. |
| Shared with Shared with {n} minds. | The card was shared with the Minds you picked. Open Manage shares to check the Shared list. |
| Share removed with This mind no longer has access to the card. | The share to that Mind was cancelled; the owner's card is not affected. |
| Cards shared with Shared {n} cards with {m} minds. / The selected cards were already shared with those minds. | The batch share finished, or every selected card already had access, so nothing new was shared. Pick other Minds or a different selection if you expected a change. |
| Cards copied with Copied {n} cards to {m} minds. | The copies were created in the target Minds. Nothing to do. |
| Copy failed with Failed to copy training cards | The copy request did not finish. Fix the cause (name conflicts in target Minds, or a card still converting or being vectorized) and copy again. |
| Converted to local card "{name}" | The local copy was created and selected, and the share was removed. Edit the copy in your Mind. |
| Name conflicts in target minds: names / A training card with the same name already exists in the target mind | Rename one of the cards, then copy or share again. |
| Remove shared card? with This card is shared with you. Removing it cancels the share to your mind - the owner's card is not affected. | Removing a shared-in card only cancels the share to your Mind. Confirm with Remove, or use Convert to local first if you need your own copy. |
| This card is shared with {names}. with It will be removed from this mind and its ownership moves to one of them - nothing is deleted. | Deleting an owned card that is still shared hands it over instead of deleting it. Continue to move the card, or remove its shares first if it really should be deleted. |
| This will permanently remove this training card from this mind. This action cannot be undone. | Deleting an owned card that no other Mind reads removes it for good. Confirm only when you are sure; there is no undo. |
| Converting a shared training card also requires the View Training Cards permission | Ask for View Training Cards. |
| Training card "name" is still being vectorized. Try again in a moment. / Training card "name" is still converting its PDF. Try again once the conversion finishes. | Wait, then try again. |
| Enabled n, skipped m (still converting or empty). | The batch enable finished; some cards were skipped as ineligible. Check the skipped cards in the Explorer; they are still converting, failed or empty. |
| Enabled training content is very large (over 400,000 chars) and may exceed the mind's context window. | The enabled persistent cards are large enough to crowd the Mind's context window. Disable the cards you do not need, or move long reference material to Vectorized cards. |
| No persistent cards in the prompt yet | The Hub footer reports no enabled persistent cards, so nothing is added to the Mind's instructions. It is a status line. |
| Some requested cards do not exist | The request named cards that have since been deleted. Refresh the Training Hub and select the cards again. |
| Remembering "{name}" with Memory created | The Mind saved a memory from the chat as a training card; the card arrives disabled. Open View in Training Hub on the chat card (needs Edit Mind), review the draft, then enable and save it. |
| Training Hub is desktop-only | The Training Hub does not run in this window; it needs a desktop-width screen. Open the Mind on a desktop screen or a wide window. |
| No training cards yet with Add your first card | This Mind has no cards yet. Click Add Card to write one, or use New from template…. |
| No matching cards or folders | The search text or the filters match nothing. Clear the search text or Reset filters. |
| Select a card to edit | No card is open in the editor panel. Pick a card in the Explorer, or create a new one. |
| Folder already exists with That path or folder path is already in the explorer. | A folder with this path is already in the tree. Choose another name, or use the existing folder. |
| Parent folder missing with Create "{parent}" first, or enter the full path (e.g. A / B). | The parent segment of a nested folder path does not exist yet. Create the parent folder first, or type the whole nested path. |
| A card with this name already exists | Card names must be unique within a Mind. Choose a different name. |
| Failed to save card | The save request did not go through (shown under the toast title Error). Try again; if it keeps failing, check the card and the connection. |
| This card was not saved. Edit the highlighted parts to remove or mask them. | The save was rejected for personal data; the flagged words are listed in the dialog. Edit or mask the highlighted parts, then save again. |
| Detecting PII… | The save is running its server-side personal-data check (non-vectorized saves). Wait for the check to finish. |
| Close tab with unsaved changes? with "{name}" has unsaved edits. Closing this tab will discard them. | The tab being closed has unsaved edits. Confirm Close tab only if the edits are not needed, otherwise keep the tab open and save. |
| Saved version | The History list shows a single fallback entry when no revisions exist; it represents the text as last saved, and Replace is hidden. Nothing to restore; older versions do not exist for this card. |
| No templates match "{query}". | The search text matches no template name, description line or tag, and the Blank card option is hidden too. Clear the Search templates... box or shorten the text; the full list returns. |
| You see | What it means and what to do |
|---|---|
| Searching knowledge | The knowledge search is still running in the open view. Wait for the passages to appear. |
| 0 passages | The search ran and matched nothing. Check that the Mind has Vectorized training cards, and review its similarity settings in AI Config. |
| No passages returned. | The search came back empty. Adjust the question, or check the Mind's training cards and similarity settings. |
| Knowledge search | The search has no query text, so the step uses this label. It is a label, not an error. |
| Search similar | The search failed and there is no query text to name it. Read the error shown with it and try the search again. |
| Untitled card | The training card behind the passage has no name. Name the card in the Training Hub, if you can edit it. |
| Open this training card in Hub | Opens the passage's training card in the Mind's Training Hub, with the card selected. If nothing happens, the passage carries no card and Mind reference and cannot be opened. |
| Chunk {i} / {n} | The passage covers part i of the card's n text chunks. Nothing to do; it shows which part of the card the passage comes from. |
| Show excerpt / Hide excerpt | Opens or closes the passage text on the card. The control is absent when the passage has no text. |
| Copy results | Copies the search results from the open view. The copied feedback resets by itself. |
| Using: {title} v{version} | The Mind loaded that skill. Open the chip to read the skill's details and text. |
| {title} v{version} already loaded | The same skill was already loaded in this turn and nothing was added again. Nothing to do. |
| A DataMind playbook shipped with this platform version | What a skill is; shown in the chip's tooltip and on the detail header. Nothing to do. |
| Loaded into the conversation | The skill's State row on a first load. Nothing to do. |
| Already in this turn, nothing added again | The skill's State row on a repeat load. Nothing to do. |
| Main text | The Reference row fallback when the skill has no reference. Nothing to do. |
| Loading the skill text | The Skill text panel is fetching the skill's text. Wait. |
| Copy text | Copies the skill's text. Nothing to do. |
| Updated to v{version} since this conversation loaded v{loadedVersion}; the model follows this text from its next request. | The skill changed after this conversation loaded it. Nothing needed; the Mind follows the new text from its next request. |
| The skill text could not be loaded. | Fetching the skill's text failed. Reopen the panel, or refresh the chat and try again. |
| You see | What it means and what to do |
|---|---|
| Benchmarks are for analytical Minds. | The Benchmarks tab is greyed out on a Warehouse Builder or Lake Builder Mind (shown in its hover tooltip on a desktop, and as the placeholder inside the tab). Run benchmarks on an Analytical Mind instead; builder Minds do not support benchmarks. |
| Questions are locked while an experiment is running on this mind | Question-bank changes are blocked while any experiment on the Mind is running. Wait for the experiment to finish, or stop it with Stop experiment. A paused experiment does not lock the bank. |
| Delete question? with This question will be removed from the bank. Question snapshots inside already-launched experiments keep their copy. | Deleting a bank question does not change experiments that already launched; they keep their own snapshot of it. Confirm with Delete if you still want the question removed from the bank. |
| This mind has no benchmark questions yet - add questions to the bank before launching an experiment. | The experiment wizard shows this warning when the question bank is empty, and Launch stays disabled. Add at least one question on the Question Bank first. |
| An experiment is already running on this mind - launching another will be rejected until it finishes. | One experiment runs at a time per Mind; a second launch is refused while one runs. Wait for the running experiment to finish, or stop it. |
| No models are available for benchmarking on this mind. | The variant picker found no model the Mind may use for benchmarking. Allow models for the Mind on its AI Config tab, or ask an administrator to allow them. |
| Another benchmark experiment is currently running elsewhere in the system - consider waiting to save resources. | A soft warning on the launch confirmation; it deliberately names no Mind and no experiment. Wait if you want to free up resources before launching. |
| Launch experiment? with This will create {n} conversations on this mind - one per question x instance x variant combination. / Every conversation consumes real model tokens, so a large experiment can burn a significant amount of usage. | The launch confirmation shows the exact number of conversations the experiment will create and warns about token use. Check the number, then continue with Launch, or cancel and reduce instances or variants. |
| In-flight items in "{name}" finish naturally and everything else pauses. All results collected so far are kept, and you can continue later. | Stop experiment pauses the run: it does not cancel in-flight items and it keeps the results. Confirm with Stop; resume later with Continue experiment. |
| Stop the experiment before deleting it | Delete experiment is disabled while the experiment is running. Stop the experiment first, then delete it. |
| PII scan found potential PII in the question bank. Edit the flagged questions or disable the scan. | With the question-bank PII scan on, the launch is rejected when personal data is found in the question or option texts; the message continues with Findings: Question N: label, .... Edit the flagged questions, or switch the scan off and launch again. |
| Benchmark in progress - the composer unlocks when this item finishes. | The open chat belongs to a benchmark item that is still queued or running, so the message box is locked. The lock lifts on its own when the item finishes. |
| Failed to load questions / Failed to load experiments / Failed to load the experiment / Failed to load the scoreboard / Failed to load per-question results / Failed to load the diagram | The panel's data could not be fetched (the server message is shown under the title when there is one). Select Try again. |
| Launch experiment failed / Add question failed / Stop experiment failed / Delete experiment failed | Read the description and try again. |
| Enter a whole number between 1 and 50 | Enter a whole number in that range. |
| You see | What it means and what to do |
|---|---|
| Name is required / Prompt is required / Mind is required | Fill in the field. |
| Select at least one workflow to watch. | Pick at least one workflow for the trigger, or turn the workflow trigger off. |
| No workflows available to watch. | There are no workflows of the kind this builder task can watch (a Lake Builder sees Lake workflows, a Warehouse Builder sees Warehouse workflows). Add a workflow in the matching area, or leave the workflow triggers off. |
| Scheduled task created | The task was created and saved. Open the task and test it with Run now, or wait for the schedule. |
| Discard unsaved changes? | You are closing the task form while it holds unsaved edits (or an uncreated task). Choose Discard to close, or Keep editing to stay. |
| Task scheduler not created / Failed to create task scheduler. | The Mind could not create the task from the chat. Read the reason on the card and create the task from the Scheduled Tasks page instead. |
| Execution requested | Run now was accepted and a run is starting. Watch the Runs card; the run appears with the Manual badge. |
| Could not create task / Could not update task | Read the description and try again. |
| Execute failed with Task is already running | The task is running. Wait for the run to finish, then select Run now again. |
| No runs yet. Use Run now or wait for the next schedule. | The task has not run yet. Start a run with Run now to test it, or wait for the schedule. |
| Only the creator of a scheduled task can edit it. | Ask the creator to make the change; an administrator can still run or delete it. |
| Only the creator of a scheduled task or an admin can run it. / ...can delete it. | Ask the creator or an administrator. |
| Restored version v{N} | The task now uses the configuration of version N. The restore is done. |
| Identical to the live configuration / This version's mind is no longer available | The version cannot be restored; it matches the live task, or its Mind is gone. Pick another version, or update the live task directly. |
| This mind doesn't include it | The task cannot use that model option because the Mind does not include the feature. Turn the feature on for the Mind in its AI Config, or leave the option off. |
| Delete scheduled task? with This will remove this scheduled task. This action cannot be undone. | The delete confirmation for a task; the removal cannot be undone. Confirm with Delete, or close the dialog to keep the task. |
| Could not restore version | Try again. |
| You see | What it means and what to do |
|---|---|
| Failed to load artefacts with Reload page | The Artifacts tab list could not load. Click Reload page; if it fails again, reload the browser tab. |
| No artefacts found with Try adjusting your filters | No artifact matches the filters you set on the Artifacts tab. Change the search, or clear the filters with Reset Filters. |
| Failed to load artifacts / Failed to load artifact / Failed to load query | Try again. If the chat's Artifacts panel is the one failing, close and reopen it, or reload the page. |
| No Artifacts yet with Artifacts from your conversation will appear here. They represent generated queries, reports, presentations, and other structured data. | This chat has not produced any artifacts yet. Ask the Mind in this chat for a query, deck or report; artifacts appear as they are made. |
| No Artifacts match your search | The search or filters exclude every artifact in this chat. Change the search or clear the filters. |
| Unknown artifact type with The artifact type "{type}" is not supported | This kind of artifact has no viewer on this page. Open it from the chat that made it, or download it from its card when Download is offered. |
| No data available / This artifact has no data | The query returned no rows when the artifact was saved. Run the query again in a chat with different filters. |
| Failed to load report / Failed to load presentation | The artifact could not be loaded. Open it again from the list. |
| Could not render preview / Preview not available with Download presentation | The preview cannot be shown in the app. Download the presentation instead, or open it again. |
| Could not download the artifact. / Download failed with Could not download this file. Please try again. | The download did not finish. Try it again. |
| {artifactType} Generated Successfully with Click to view the artifact | An artifact finished building in the chat. Click View, or open it from the artifact list. |
| type Generation Failed with An error occurred during generation | For example, Presentation Generation Failed. The artifact did not finish generating. Open View to see the step, then ask the Mind to try again. |
| Could not open this conversation. | The conversation linked from the answer cannot be opened. It may have been deleted, or you may not have access to it. Ask for access to that chat, or keep working from the artifact itself. |
| Could not download the presentation. | Try the download again. |
| Could not create the training card. | The card from Train based on this query (or from a template pick) was not created. Try again; if the SQL may hold personal data, review it first. When the server sent a reason, the toast shows it instead of this text. |
| A table with this name already exists | Choose another name, or select Open existing. |
| Failed to create view / At least one logical model is required / Minimum 2 characters required | Fix the highlighted field and try again. |
| Not enough content in conversation to generate artifacts | The chat is too short to build an artifact. Continue the conversation, then ask for the artifact again. |
| Required permission(s): View Artifacts OR View Others Artifacts | Your account cannot open the Artifacts routes. Ask an administrator for View Artifacts (your own artifacts) or View Others Artifacts (everyone's). |
| No themes found with Try different search or filters | No theme in the catalog matches the search text or the selected category chips. Clear the search box or click a chip again to remove it; the themes return. |
| You see | What it means and what to do |
|---|---|
| Approval failed / Rejection failed | The decision on the proposed change did not save. Read the description and try again. |
| Approving a destructive change requires the Warehouse modify permission | A destructive warehouse change can only be approved with Add/Modify Warehouse Objects, so the decision stays locked for users without it. Ask an administrator for the permission, or reject the change on its card. |
| Approving a destructive lake change requires the Lake modify permission | A destructive lake change can only be approved with Add/Modify Lake Objects. Ask an administrator for the permission, or reject the change on its card. |
| Changes cannot be approved from an impersonated session. | A decision was attempted while viewing DataMind OS as another user. Exit the view, then approve or reject on the card. |
| A decision is already being applied to this conversation. | A second decision arrived while the first was still being applied. Wait for the first decision to finish; the card updates with its outcome. |
| A response is still being generated for this conversation. Stop it before deciding. | A decision was attempted while the Mind was still answering. Wait for the answer to finish, or stop the response from its control, then decide. |
| Approval pending with Approve or reject the proposed warehouse change above before sending a new message. | A proposed change waits for a decision, so sending is paused. Approve or reject it on its card, and the message box opens again. See Builder changes. |
| Rejected by you / Rejected by another admin / Rejected by an admin (with the written reason after it when one was given) | The card names who rejected the change. The call was not executed, and it cannot be approved afterwards. |
| Nothing was executed for this call. | The rejected call never ran. Rejection is final for that call; if you still want the change, let the Mind propose it again. |
| You see | What it means and what to do |
|---|---|
| Every warehouse change waits for your approval / Every lake change waits for your approval | The Manual mode description in the change acceptance picker. Pick the mode you want; in Manual, decide each approval card as it appears. |
| Changes apply automatically; drops and deletes still ask | The Auto-accept mode description; ordinary changes run on their own and destructive ones still wait. Watch for the destructive approval card. |
| Everything runs without asking, including drops and deletes. Requires the Warehouse remove permission. / Everything runs without asking, including drops and deletes. Requires the Lake remove permission. | The Unrestricted mode description and the right it needs. Hold Remove Warehouse Objects (or Remove Lake Objects) to switch it on. |
| Requires the Warehouse modify permission / Requires the Lake modify permission | The tooltip on the greyed-out Auto-accept option. Ask an administrator for the right, or stay in Manual. |
| Requires the Warehouse remove permission / Requires the Lake remove permission | The tooltip on the greyed-out Unrestricted option. Ask an administrator for the right; Unrestricted cannot be set without it. |
| Unrestricted mode can only be switched on per chat, by users with the Warehouse remove permission. / Unrestricted mode can only be switched on per chat, by users with the Lake remove permission. | The note under Default change acceptance; Unrestricted is never a stored Mind default. Switch it on per chat, or ask an administrator who holds the remove right. |
| Enable Unrestricted mode? | The confirmation before Unrestricted turns on; destructive actions will run without asking. Confirm with Enable Unrestricted mode, or cancel to stay in the current mode. |
| Unrestricted mode: destructive warehouse changes run without approval / Unrestricted mode: destructive lake changes run without approval | The stripe shown while Unrestricted is on. Use Switch to Auto-accept to end it. |
| Applies to tool calls evaluated after the change. Approvals already waiting stay pending. | The footer of the change acceptance picker; changing the mode does not release or cancel anything already waiting. Decide the waiting cards themselves. |
| Applying the new mode... | The chip is locked while a mode change is being saved. Wait for the save to finish. |
| Change acceptance mode is controlled by the root chat. | The server refusal when a child chat tries to set its own mode. Set the mode on the root chat. |
| Auto-accept mode requires the Warehouse modify permission. / Auto-accept mode requires the Lake modify permission. | The server refusal when raising the mode without the matching modify right. Hold Add/Modify Warehouse Objects (or Add/Modify Lake Objects), or stay in Manual. |
| Unrestricted mode requires the Warehouse remove permission. / Unrestricted mode requires the Lake remove permission. | The server refusal when raising the mode without the matching remove right. Hold Remove Warehouse Objects (or Remove Lake Objects), or use Auto-accept. |
| Set on the root chat of this swarm | The change acceptance chip is locked in a swarm agent's chat; the child does not set its own acceptance mode. Change it on the main conversation. |
| You see | What it means and what to do |
|---|---|
| This tool cannot be stopped from its card. | The step is not one that carries stop controls; only certain steps can be stopped or aborted from their card. Let the step finish, or end the job it belongs to with Abort run if that button appears. |
| This tool call is not running. | The step already finished, failed or was stopped, so there is nothing left to stop. The result in the chat stands; Abort run can still end a live job that the step started. |
| Stopping a tool or reading its run status in this conversation requires access to Minds chats or the data layer. | Your access to this conversation does not cover its tool controls. Ask an administrator about access to Minds chats or the data layer. |
| Aborting an ETL run requires the Warehouse modify permission. | The role lacks Add/Modify Warehouse Objects, so it cannot abort the ETL job. Ask an administrator for the permission. |
| ETL job {id} is not running. | The ETL job finished before the abort request arrived. Nothing to do; the run ended on its own. |
| Aborting a workflow run requires the Execute Workflows (Warehouse) permission. | The role lacks Execute for Warehouse workflows, so it cannot abort the workflow run. Ask an administrator for Execute Warehouse Workflow Objects. |
| Aborting a lake run requires the Execute Lake Workflow Objects permission. | The role lacks Execute for Lake workflows, so it cannot abort the lake run. Ask an administrator for Execute Lake Workflow Objects. |
| Workflow run {id} is not running. | The workflow run finished before the abort request arrived. Nothing to do; the run ended on its own. |
| Stop requested | The stop request was accepted; the step stops from the chat's point of view while work on the server keeps running. Wait for the step's result, or abort the run if the job itself should stop. |
| Workflow run {jobId} aborted | The workflow run was stopped and its open steps are marked Aborted. Nothing to do; restart the run from the Workflows page if it should run again. |
| Run aborted / Run aborted (synchronize days restored to {n}) | The ETL run was stopped where it was; with a truncate-insert pattern the table can be left partially loaded, and a backfill's earlier synchronize days are restored. Check the table before using it, and run the load again if it should complete. |
| Stop failed / Abort failed | The stop or abort request did not go through and the run keeps going. Read the reason under the toast, then try again from the card. |
| You see | What it means and what to do |
|---|---|
| Stop failed / Continue failed / Start again failed | The action on the Model Builder session did not save. Read the description and try again from the session card; if it keeps failing, reload the chat. |
| Needs you | The Model Builder is waiting for a decision from you, a question or an approval. Open the session card and act on it. |
| Model Builder · {model} · {round text} · {state} | The session strip above the message box shows the live session; its badge reads Needs you: {n} when questions or decisions wait. Select the strip to open the session card. |
| Model AI building, round {n} of {m} | A round of the build is running in the model's Model AI chat. Nothing to do; follow the round dots on the session card and the strip. |
| Waiting for the Model AI chat / Round {n}: waiting for the Model AI chat to be free | The round is parked because the Model AI chat is busy with other work. Wait; the round starts by itself when the chat is free. |
| Reading the live status | The card is loading the session's live status. The card fills in when the status arrives. |
| The live status of this Model Builder is not available to you right now. | The card could not read the session's live status, so it shows the last stored status instead. Open the card again later; the live view returns when the session can be read again. |
| The Model Builder stops working on this model. Anything it made stays; you can start it again. | The stop confirmation text; stopping keeps everything already built. Select Stop in the dialog to end the session now, or cancel; start it again later from the card. |
| Only {name} can give it 2 more rounds. | Only the person who started the build can add the extra rounds. Ask that person to select Continue with 2 more rounds on the session card. |
| The Model Builder needs a change on a shared table | The build wants to change a table that existed before the design; it changes only the tables the design created or changed. Choose I will make it in the Warehouse Builder chat to do the change yourself (the build skips it) or Continue without it to carry on without the change. |
| This table existed before this design, so the Model Builder changes it only with your word. | The hint under the shared-table question; older tables are changed only once a person allows it. Decide between the two options in the question. |
| Someone who can write in this chat answers this question. | The question is for people who can write in the chat. Answer it yourself if you can write there, or ask someone who can. |
| Send anyway | Your answer tripped a personal-data warning and can be sent on with the warning bypassed. Select it to send the answer as it is; an answer the check blocks (rather than warns about) cannot be sent on. |
| Model Builder stopped | The stop you asked for succeeded. The session is closed and can be started again. |
| Model Builder sessions reset | Start again reset the model's session count. Follow the new session on its card. |
| Answer recorded | Your answer was saved and the build can continue. |
| Used all 8 rounds. Continue with 2 more rounds, or open the Model AI chat to finish by hand. | The session used all its rounds. The person who started the build can add 2 more rounds, or finish by hand in the Model AI chat. |
| Reached its 12-hour limit. Start a new session from this chat, or finish in the Model AI chat. | The session reached its time window. Start a new session, or finish by hand in the Model AI chat. |
| Used its 30M token budget. Start a new session from this chat. | The session used its token budget. Start a new session from the chat. |
| The Model AI asked for the same thing three times. Decide here, then start a new session. | A repeated request stopped the build and an item is waiting for a decision. Settle the item in Needs you, then start a new session. |
| Two rounds in a row did not finish. Start a new session. | Rounds kept failing to finish. Start a new session from the chat. |
| The requester can no longer use this model's Model AI. | The person who started the build lost access to the model's Model AI. Ask for the access back, then start a new session. |
| The model was deleted. | The semantic model this session worked on was removed. Start over with a new design if you still need the model. |
| Finished without a complete model: see the open items. | The build ended while items remain. Read the open items on the done card and finish the rest in the Model AI chat. |
| The Model Builder could not start. / The Model Builder failed while it worked. | The session did not start, or it crashed mid-run. Start a new session from the chat; if it keeps failing, check the model in DataPrism before retrying. |
| A Model Builder is still working on this model. | Another start was refused because a session is already open on the model. Wait for the open session to end, then start again. |
| Only people who can use this model's Model AI can read it. | The full Model AI text (a question, a change request or a report) is shown only to people who can use that model's Model AI; others see a stub. Ask someone with Model AI access, or ask an administrator for the permission. |
| Slow: recorded as an open item | The chart's timing check marked it as too slow, and it was recorded as an open item. Review the chart under Chart timings on the report and in the done card's open items. |
| The query settings could not be restored: {error} | The backfill load could not restore the saved query settings and runs with the eco settings merged over the custom ones. Note the error text; the load continues. |
| You see | What it means and what to do |
|---|---|
| Needs your approval | The design is waiting for your decision; the model is not built until you decide. Read the design sections, set the choices, then approve it or reject it with a note. |
| Row-level security will not apply with This model ignores the warehouse row-level security of {rlsTableName} ({reader}): every report viewer sees every row unless row-level security is added on the model. | An admin owns the design and the source table's warehouse row-level security does not carry over to the model. Tick I understand to approve, or add row-level security on the model to restrict rows. |
| Row-level security will not apply with {objectName} does not carry the row-level security of {rlsTableName}, so {roleName} and every report viewer of this model will see every row. Fix: hand over {rlsTableName} itself, or add the same rule in {objectName}'s RLS tab. | The business-owner form of the same warning; the fix is named in the text. Hand over the source table itself, add the same rule in the object's RLS tab, or tick I understand to approve. |
| None of the requester's roles can own a semantic model. Reject it with a note. | No role the requester holds can own the design, so approval is blocked. Reject with a note; the Warehouse Builder proposes a corrected design. |
| Pick another owner role: admin needs views over every table of this design. | The admin role cannot own this design because it lacks views over every table. Pick another owner role on the card. |
| Tick "I understand" under the row-level security warning to approve. | The row-level security warning is not acknowledged yet. Tick I understand on the card. |
| The design card could not be built for this proposal, so only its summary is shown. | The full design could not be rendered, so only the summary is available (the note adds that approving applies the choices from the intake, and that a design reading tables with row-level security cannot be approved this way). Approve to apply the intake choices, or reject with a note so the Warehouse Builder proposes it again. |
| Approved by {who} · {date} / Rejected by {who} · {date} | The design was decided; the name is the decision maker ("you" when it was you, "an admin" when an administrator decided). Nothing needed; the model is created only when the requester continues the conversation. |
| Rejection note: {reason} | The design was rejected and the note you wrote stays on the card. The Warehouse Builder reads the note and can propose a corrected design. |
| Grant {roleName} access | The access check found objects the model's owner role cannot read yet, so the check waits for you to grant them. Grant each listed object, then select I've granted them, check again; the card shows its done title when everything is granted. |
| Open each object below, add {roleName} under Permission Groups, then Save. Tell me when you are done. | The card instructs you how to grant access for each object before the model can be created. Open each object's permissions, add the role under Permission Groups, save, then select I've granted them, check again. |
| I granted the access. Check again. | The message the re-check button sends into the chat; the platform checks every object again. Watch the card's progress update with the new grant state. |
| {roleName} will see in clear in SQL: {column} ({piiDomain}), … | The grant would let the role read these personal data columns in clear when it runs SQL itself. Decide whether the role should see them; the warning is information for your grant decision, not a blocker. |
| {roleName} will also see: {columns} | The grant would expose further columns to the role. Review them before granting; adjust the design with the builder if the role should not see them. |
| Has its own row-level security: {roleName} reads only the rows its rule allows. | This object applies its own row rule, so the role's reads through the model stay limited by it. Nothing required; this is information about the object's protection. |
| {column} is not marked as personal data: reports would show it in clear too. The Warehouse Builder must mark it before the model can be created. | A blocker; the model cannot be created until the column is marked as personal data in the Warehouse. Ask the Warehouse Builder to mark the column, then select I've granted them, check again. |
| (not marked) | Shown after a column in the role access details: the column is not marked as personal data, so reports would show it in clear. The Warehouse Builder must mark it before the model can be created. |
| Created empty in the requester's personal workspace; move it later with Move. | The new model has no content yet and lives in the requester's personal workspace. Open the model and move it with Move when you know where it should live. |
| If its personal data check flags the brief, the Model AI panel shows what it found before anything is sent. | The brief is checked for personal data, and anything it finds is shown to you in the Model AI panel before the Model AI receives the brief. Review the findings, then continue with the build. |
See Swarm agents for the tree, the statuses and the reports.
| You see | What it means and what to do |
|---|---|
| Failed to stop agent | The request to stop one swarm agent did not go through. Try again from the swarm tree; if it keeps failing, reload the chat and retry. |
| Failed to stop swarm | The Stop all request did not go through, so live agents are still running. Try Stop all again. |
| Failed to resume agent | The request to resume a stopped or failed agent did not go through (the server's reason appears with the message when it sends one). Check the agent's state in the tree, then try Resume again. |
| Message not delivered / Agent not stopped / Agent not resumed | A card action on an agent did not go through. Try the action again; the card shows the agent's current state. |
| Swarm agents not spawned / Failed to spawn swarm agents. | The spawn step failed, so no sub-agents started for that request. Review the step card, then send the request again. |
| Swarm status unavailable / Failed to check swarm status. | The status check failed. Try the step again, or open the swarm tree for the live state of the agents. |
| Swarm report failed / Failed to file the swarm report. | An agent's report was not filed, so the run's report may be incomplete. Try again; if it keeps failing, ask the agent to summarise its work in its own chat. |
| No swarm agents yet | The tree has no agents, because swarm has not run in this chat. Turn on the swarm toggle in the composer and ask the Mind to parallelize its work. |
| Visibility is controlled by the parent chat | An agent's chat follows the main conversation's visibility, and the switch is locked. Change the visibility on the main conversation. |
| A swarm agent needs your decision (or the agent's name, then needs your decision) | An agent is blocked until you answer it or approve something. Open the agent's chat and answer there; the notice's Open button takes you to it. |
| No report details available. | The swarm report has no content to show for this run. Open the run's swarm entries in Tool History, or the agent's chat, for what was recorded. |
| Artifact not available | An artifact named in a swarm report cannot be opened because it could not be resolved. Open the run's artifacts from the chat's Artifacts panel instead. |
| Result has 150+ rows - the agent executes the query itself instead. | The parent delivered the artifact without result rows because the result is too large. Nothing to do; the agent runs the query itself. |
| The query was never executed - there were no stored result rows to attach. | No result rows existed when the artifact was delivered to the agent. Nothing to do; the agent runs the query itself. |
| Result rows are too wide to fit the attachment budget - the agent executes the query itself instead. | The parent delivered the artifact without result rows because the rows are too wide. Nothing to do; the agent runs the query itself. |
| No artifact details available. | The Delivered artifacts card in an agent's chat has no entries to show. Check the agent's own chat steps for what it did. |
| You see | What it means and what to do |
|---|---|
| Workflow not available / It may have been deleted, or you do not have access to it. | The workflow behind a chat chip cannot be opened; a deleted workflow and one you cannot read show the same notice. Check the Workflows screen to see whether the workflow still exists and is available. |
| Workflows could not be loaded / Close this window and open the workflow again. | The window could not load the workflow list it needs for the jobs or the diagram. Close the window and open the workflow again from its chip. |
| No Warehouse workflows. | The workflow list is empty. Nothing to fix; the step returned no workflows to list. |
| No finalized steps - nothing is in the plan. | None of the workflow's transformations are finalized, so the plan has no steps. Finalize the transformations that belong in the plan. |
| not finalized - not in the plan | A transformation is excluded from the workflow plan because it is not finalized. Finalize it, then it can take part in the plan. |
| No runs recorded. | The workflow has no runs to list. Nothing to do; run the workflow when needed. |
| No steps recorded for this run. | The run's step list is empty. Nothing to do; reopen the run later if you expect steps. |
| No error text recorded. | A failed step carries no error text. Nothing to do; read the step's status and the run's log steps. |
| run id not resolved yet | A started run's id was not known when the card was drawn. Find the run on the Workflows screen. |
| Enabling or disabling a workflow happens in the Workflows page, never from the chat. | Chat cannot switch a workflow on or off by design. Change the workflow on the Workflows screen. |
| Listing Warehouse workflows… | The workflow list card is loading. Wait; the list replaces the subtitle. |
| Reading the workflow plan… | The plan card is loading the workflow's steps. Wait; the plan replaces the subtitle. |
| You see | What it means and what to do |
|---|---|
| No structured result. | The step finished but its result was not in a shape the card can lay out. Open the step in Tool History to see what came back; ask the Mind again if you expected data. |
| Nothing to show. | A table came back empty. Open the step in Tool History; check the step's arguments if you expected rows. |
| None. | The dependants list is empty. Nothing to fix; there are no dependants of that kind. |
| No columns checked. | The column health check found no columns to check. Check that the step ran against the right table. |
| No partitions. | The partitions list is empty for this table. Check the table's partitioning; a table without partitioning says Partitions are not supported for this table. |
| Partitions are not supported for this table. | The table has no partitions. Nothing to do; partition views do not apply to it. |
| No queries matched. | No queries fit the applied filters or window. Widen the window; the Filters section lists what was applied. |
| No tables were checked. | The unused-tables check returned nothing to judge. Re-run against the right database scope if you expected results. |
| No active parts matched. | The partition stats found no active parts for the filter. Widen the filter or check again later. |
| No ETL runs recorded. | There is nothing to list for this transformation or table. Check the object in the Warehouse. |
| No refreshable view is deployed yet. | There is no view refresh state to show. Deploy the view first. |
| No warehouse objects were created or updated in this window. | Nothing changed in the period the change log covers. Widen the window. |
| The tool cut the list at its row limit | Shown as the hover text of the Truncated chip; the tool returned only part of the list. Ask for a narrower period or a smaller object. |
| +{N} more - open the tool view for the full list. | The card shows only part of a longer list. Open the step in Tool History for the full list. |
| +{N} hidden | A dependants list is longer than the card shows. Open the tool view for the rest. |
| Showing {X} of {Y} partitions. | The partitions table shows a preview. Open the tool view for the full list. |
| Running… | The step is still working (the cards use their own variants, such as Reading the query plan… and Aggregating table health…). Wait; the card shows a result when the step finishes. |
| Still running: the agent will keep waiting / Wait timed out, still running | The wait gave up but the job continues. Watch the run from its card; abort it if you may (see Stopping runs and tools). |
| Waiting for the ETL engine | The run started but the engine has not reported progress yet. Wait; the progress line updates when details arrive. |
| Refused / Blocked | The step did not run and the reason shows in place of its summary; the full view prefixes the message with Refused: or Blocked:. Read the reason and adjust the request. |
| Deploy failed for table {id} | The deploy attempt did not succeed. Read the error on the card and retry, or fix the cause. |
| Could not write for the loaded data: {error} | Writing the index data for the table failed. Read the error; redeploy after fixing it. |
| Not needed: the table is empty | Nothing had to be written for the indexes. Nothing to do. |
| Whole table truncated. | Every partition was truncated. Reload the table if you need the data back. |
| Column "{name}" was dropped | A column and its data were dropped. Check the objects that read this table. |
| ETL run history rows deleted: {N} | The truncation also removed run history. Nothing to do; the count tells you how much was removed. |
| No candidate skips at least 25% of the kept granules | No index candidate would skip enough granules to help. Nothing to do; the queries work as they are. |
| Needs ClickHouse {version} | The index estimate needs a newer database version than the server runs. Use the query plan step meanwhile. |
| Woken up after {duration} / Slept {duration} / Woken early | The sleep ended early or after the planned time. Nothing to do. |
| Swapped in one step: readers kept the view until the table was filled / Dropped and recreated: no rows until the first refresh finished | How the refreshing view replaced the previous one. Nothing to do for the swap in one step; for drop-create expect no rows until the first refresh finishes. |
These appear on the cards under an answer, or in Tool History. See Tool cards in chat and Charts in answers.
| You see | What it means and what to do |
|---|---|
| Tool call not found in this conversation. | The step's details cannot be found in this chat, for example after a reload. Open the step again from its message; if it stays missing, ask the Mind to repeat it. |
| The tool call failed. | The step did not finish. Open the card to read the reason, then ask the Mind to try again. |
| Couldn't display this content. / Couldn't display this tool result with Try again | One block of the answer could not be drawn. Click Try again; the rest of the answer stays readable. The label names the failing block, for example a message or an artifact. Mid-stream a broken step draws nothing and the notice stays only if it is still broken after streaming ends. |
| Running… / Done | A step's card while it works, and when it finished with its result inside. Open the step in full if the summary is not enough. |
| Success / Error / Awaiting approval / Rejected / Skipped / Pending | The step's status chip: it finished; it failed; it is held until you approve or reject it; a held step was rejected; it did not run; it has not started. For Awaiting approval act on the card; for Error open the step to read why. |
| No tools available | There are no steps in this chat to list for the tool filter. Close the filter; steps appear in the panel as the Mind works. |
| No tools match your filters | The search or the tool filter hides every step. Change the search term, choose All Tools, or select Reset. |
| Plan update rejected / Failed to update the plan. | The Mind's plan change was not applied. The previous plan stays; ask the Mind to try again if the plan matters. |
| 0 rows / {n} out of {m} rows / {start} - {end} out of {total} rows | The query came back empty, or a limit or a page cut the result. Open the step in full for the rest; export once the result has more than 50 rows. |
| Fetching databases... / No databases found / No tables found | The catalogue step is loading, or nothing came back for it. Wait for the fetch; if the result is empty, check the database name or ask the Mind again. |
| Initializing... / AI is searching the web... | The web search step is preparing or running. Wait; the results appear when the search finishes. |
| Unable to render chart / The chart data could not be displayed with Copy details | One chart failed to draw; the rest of the answer stays readable. Copy the details for support, then ask the Mind for the chart again. |
| Invalid chart configuration with Show raw code | The chart's settings cannot be drawn. Open the raw code to inspect it, then ask the Mind to fix the chart. |
| Generation was interrupted. Showing the {n} rows received | The answer stopped mid-stream and the chart drew from the rows that arrived, with the count of what is shown (a variant adds , {k} could not be read). Ask the Mind again if the missing rows matter. |
| Showing {n} of {m} rows. {k} could not be read. | Part of the chart data failed to parse; the readable rows are drawn and the unreadable ones are dropped. Ask the Mind again when the full set is needed. |
| Part of the chart data could not be read and is not shown. | Unreadable text between data entries; counts are deliberately not shown because they would read as complete. Ask the Mind again for a complete chart. |
| Chart data is corrupted: the {field} field could not be read | A required field of the chart data could not be parsed. Ask the Mind again; use the raw-code toggle to see what arrived. |
| Unsupported chart type: {chartType} | The answer used a kind of visual that chat cannot draw. Ask for one of the supported kinds (pie, bar, line, area, combo, radar, radial bar, Sankey, or a diagram). |
| No code available | The code view opened but this chart block carries no settings to show. Nothing to fix; the chart and table views still work. |
| Loading diagram / Generating | A diagram block is waiting for its rendering engine to load, or the answer is still being written and the chart mounts when the answer completes. Wait a moment; the drawing replaces the label. |
| Chart generation interrupted / Chart generation was interrupted before it could be read with Show code | The block was cut off before it finished. Open the code, then ask the Mind again. |
| Invalid diagram syntax / Diagram Error / Could not validate diagram syntax | A diagram could not be drawn. Open its source and inspect it, then ask the Mind again. |
| Problem found on line {n} | The line of the diagram's code where reading stopped. Read that line in the view source, then ask for a corrected version. |
| Diagram engine did not load with Reload page | The diagram renderer did not start. Reload the page; the diagram itself is fine. |
| You see | What it means and what to do |
|---|---|
| Running / Done / Failed / Skipped | The step's status pill: it is still going (its icon spins and its title shimmers); it finished; it failed and the card turns red; it did not run. Open the card's sections to read what it returned, and read the Error section when the card shows one. |
| Tool call | A step arrived without a name, so the plain fallback card carries this label. Read its Input, Output and Error sections; ask the assistant what the step did if it stays unclear. |
| Error | The plain card's section that holds the step's error text; it opens by default. Read it to see why the step failed. |
| No query | The web search step carried no query text to show. Open the card for the results; ask the assistant what was searched if you need to know. |
| {filePath} · Replacing... | The replacement is still being written. Wait; the line turns into the done recap with the line counts once the write settles. |
| (empty) | The old or the new side of a replace preview had no text. Read the other side for what changed. |
| · case-insensitive | A suffix the step shows when the search ignored the difference between capital and small letters, so both spellings counted as matches. Nothing to fix; this only reports how the search ran. |
| No change: {summary} | The step finished without changing anything, and the summary after the prefix says what it returned. It is not an error; the amber colour marks that no change was made. |
| Refused | The step failed, and this payload opens with the step's error text. Read the error text to see why the step did not go through. |
| Sandbox unavailable: {reason} | The preview could not run in the sandbox and the reason follows the colon. Address the reason with the assistant, then preview again. |
| Render failed | The preview ran and the frame did not render cleanly; the report lists the problems. Read the Report payload, which opens with the error lines and any blocked addresses. |
| Applied; the frame has not reported yet. | The change was applied but no render report has come back. Wait a moment; the report appears on the card when it arrives. |
| paints differ | The exported preview painted differently across runs, so the visual is not deterministic. If you need identical exports, ask the assistant to review the code for run-dependent drawing. |
| export: {error} | The export attempt failed and the error follows the prefix. Read the error; the assistant can adjust the visual and export again. |
| {n} more not shown | The findings list was cut short and n findings are not shown. Ask the assistant about the remaining findings when they matter. |
| Download successful with Diagram exported as SVG | The diagram was saved as an SVG file named after its kind. Nothing to do; the file is in your downloads. |
| Download failed with Diagram is not ready to download yet | The export ran before the diagram had finished drawing. Wait for the drawing, then click Download SVG again. |
| You see | What it means and what to do |
|---|---|
| Running the pipeline - this can take a while | The sync run is in progress; the card blocks until the run finishes. Let it finish before reading the data; the note under it says the landing database is rebuilt from scratch on every run. |
| The landing database is rebuilt from scratch on every run. | The note a running sync carries: each sync run rebuilds the connector's loaded data from scratch. Nothing to do. |
| Pipeline failed | The sync run did not complete; the card shows the error. Fix what the error names, then run the sync again. |
| Could not list tables | The connector's landed tables could not be read. Try again; if it persists, check the connector's last sync. |
| No tables - the tap produced no streams | The last sync landed nothing because the connector produced no streams. Check the connector's streams and run a sync. |
| Row counts are a sample - each stream is capped during sync. | The counts on the card are capped samples, not exact totals. Treat them as a rough size only. |
| Sample failed | The sample read did not go through. Read the error block; fix what it names, then sample again. |
| The sample hit the limit - the stream has more rows. | The preview shows only the first rows of the sample; more rows exist. Use Copy JSON for the full sampled row set. |
| Validation failed · Invalid config | The meltano.yml schema check failed. Read the error block; fix the config and test again. |
| Validation failed (on Validate recipe) | The recipe has fatal or error issues; the subtitle counts them. Fix the listed issues, then validate again. |
| Unapproved hosts: {list} | The recipe calls hosts that are not approved yet, and loads can call only approved hosts. Approve the hosts under Hosts. |
| Connection failed | The live connection check did not complete. Read the error block; check the connection settings and the approved hosts. |
| {n} host(s) still pending approval | The approval left hosts waiting. Approve the remaining hosts; the connection step cannot move on until every host is approved or revoked. |
| Request failed (on an HTTP request) | The request the assistant ran did not complete. Read the error; check the API and the approved hosts. |
| Response truncated - showing the first 32 KB | The response body is larger than the shown part; the card adds the body's full size. Read the rest from the API itself; request headers are deliberately never shown. |
| Intermittent - neither half reproduced | The page break is not stable; splitting the page did not reproduce it. Review the steps table; the break may pass on a retry. |
| Bisect failed | The search could not produce a verdict. Read the card's steps and error, then try again. |
| Probe failed | The page probe did not complete. Read the error block; fix what it names. |
| The record path is not present in this page - the loader would read 0 records from it. | The page has no records at the recipe's record path. Check the record path in the recipe, then probe again. |
| No diagnostics | Nothing could be read from the connector's Sync Log; the error under the subtitle says why. Address the error and run it again. |
| No runner diagnostics were collected for this run. | The run finished without collecting diagnostics. Read the run's status and error instead. |
| Waiting for secret(s) | The connector needs secret values before the step can continue; the vault opens once for the requested names. Store the values in the vault. |
| Secrets not provided | At least one requested secret value is still missing. Store the missing values in the Secrets Vault. |
| Could not request secrets | The secrets request failed. Read the error text on the card. |
| No data access grants yet. An admin adds one in the recipe editor. | The connector has no grants. Ask an admin to add one in the recipe editor's Data access section. |
| Waiting for an admin… | The data access request waits for an admin. Wait; an admin opens it with Open Data access and grants it. |
| No grant created | The request settled without a grant. An admin opens Open data access on the card and grants it. |
| Preview failed | The grant preview did not complete. Read the error; try again. |
The assistant panel runs beside the Lake connector and other screens. See The AI assistant.
| You see | What it means and what to do |
|---|---|
| Loading | The panel is loading the conversation list for this surface. Wait; the panel then opens your most recent conversation for that surface, or its starting screen. |
| Connecting to assistant… | The assistant panel is not connected yet, so the composer cannot be used. Wait for the connection; when it is ready the box reads Ask the assistant…. |
| Could not load conversations. with Retry | The assistant could not load your conversations. Click Retry. |
| {name or email} is running a turn... | Another person is running a turn in the conversation, so the message box is disabled. Wait for their turn to finish; the box opens again. |
| Upload failed with "{file.name}" could not be uploaded. Please try again. | A file staged for an assistant message could not be uploaded. Attach the file again to retry it; the message can still be sent without it. |
| Connection lost with Trying to reconnect. Sending is paused until the connection is back. | The assistant's live connection dropped; the badge can read Connection lost: {reason} when the cause is known. Wait; the panel reconnects on its own and sending unlocks when the connection is back. |
| Socket connection timeout / Failed to create socket instance / Failed to connect to chat server / Failed to reconnect | The assistant chat connection could not open or reconnect. Wait for the panel to reconnect; sending unlocks when the connection is back. |
| Approval pending with Approve or reject the proposed change above before sending the next message. | A proposed change waits for a decision. Approve or reject it on its card, and the message box opens again. See Builder changes. |
| Connector change needs your approval | A connector change was held and waits for your decision before anything is written. Read the rows and select Approve or Reject on the card. |
| Awaiting approval | The row's change is held and not applied yet. Decide on the card; nothing is written until you approve. |
| Approval failed / Rejection failed | The approval or rejection decision did not save. Read the description and decide again. |
| Rejected | The held change was rejected and not applied. Nothing to fix; the assistant reads the reason you wrote. |
| Why? (optional; the model reads this) | The optional reason field for a rejection. Write a reason up to 500 characters, or confirm without one. |
| Could not change the acceptance mode | The new mode was not saved; the toast adds The assistant is still running in the previous mode. and the picker returns to the mode you had. Select the mode again; if it keeps failing, check the panel's connection. |
| Available once the chat has started | This chat has no conversation yet, so there is nothing to change; the picker is greyed out with this tooltip. Send your first message, then set the mode. |
| Compacting conversation… | The assistant is summarising the older turns of this conversation to free up room. Wait; when it finishes, a Conversation compacted marker appears. |
| Conversation compacted | Older turns were replaced by a summary, so the assistant keeps the gist of them from that point on. Nothing needed; keep chatting. |
| Context cleared | The context was cleared; your messages stay visible, but nothing above this marker reaches the model. Nothing needed; the next reply starts from a blank slate. |
| Nothing to compact with Open a conversation with some history first. | There is no conversation with history to summarise. Open a conversation, or chat for a few turns, then run the command again. |
| Nothing to clear with Open a conversation with some history first. | There is no conversation with history to clear. Open a conversation, or chat for a few turns, then run the command again. |
| Compaction skipped | The compaction did not run, so nothing was summarised. Send the command again. |
| Compaction failed | The compaction did not finish. Send the command again. |
| Context not cleared | The clear did not go through, so the conversation was left as it was. Send the command again. |
| Not sent | A failed turn left this message in the assistant conversation. Click Retry to run the turn again, Continue when part of the answer arrived, or Send anyway on a personal-data warning when the bypass is available. |
| Retried | The panel is replaying the failed message after you clicked Retry; the card disappears once the new run starts streaming. If it does not start, a Retry failed toast appears and the card stays. |
| Retry failed / Continue failed | The retry of a failed turn, or the continue of a partial answer, did not start. Try the action again; the card stays available. |
| Send anyway | The bypass button offered on a personal-data warning for the assistant; shown only with the bypass ability, while the message is still the exact warned text. Click it to send the warned message as it is, or edit the personal data out and send again. |
| Detecting PII | Your message is being checked for personal data before it is sent. Wait for the check; if something needs a decision, the panel shows it. |
| Responding | The assistant is answering. Wait for the reply to finish. |
| No active stream to cancel | Cancel (Stop) was pressed when no answer was running server-side. Nothing was stopped; the message box is ready for your next message. |
| This conversation has no messages with The last message was not accepted, so nothing was saved here. Your text is in the box below - edit it and send it again. | The conversation's first turn was rejected, so the thread stayed empty. Edit your text in the message box and send it again. |
| Start a conversation to chat with your assistant. | The thread has nothing to render yet. Type in the message box to begin. |
| Asking question… | The assistant is writing the question, or it waits above the message box for your answer. Answer it in the panel above the message box, or skip it with Cancel. |
| Assistant needs your input | A question row that was never answered and never skipped; the assistant did not get what it asked for. If the panel is not open any more, send your answer as a normal message in your own words. |
| Skipped with Question skipped | The question was closed without an answer. If you skipped it by mistake, send the answer as a normal message. |
| Question not delivered with The assistant's question could not be delivered. You can reply in your own words instead. | The question's tool call failed; the choices it would have offered are still listed. Send what the assistant asked about as a normal message in your own words. |
| Could not cancel question | The skip request failed; the toast carries the platform's message. Try again, or answer the question instead. |
| Could not open this conversation. | The conversation behind a message's conversation chip did not open; the chip opens it only when your access allows it. Check whether you can open that conversation from the Minds chat list. |
| Could not load the version history. | The connector file history did not load. Close the drawer and open it again from History. |
| Loading history | The drawer is loading this connector's version list. Wait; it resolves when the connector's conversation loads. |
| No versions yet with A version is saved every time the AI assistant changes the connector files. Ask it to edit something and the version will appear here. | No version has been saved for this connector yet. Ask the assistant to edit the connector's files; the version appears here afterwards. |
| No file changes were recorded for this version. | This turn was saved as a version without file edits, so there is nothing to preview. Check the versions around it. |
| The code base was reverted to a point that was before the message below / The code base was restored to the point of the message below | The assistant rolled the connector files back, or restored them, at this point. Nothing to fix; a hover shows who did it and when, the quote jumps to that message, and Undo this revert appears while the target stays rolled back. |
| Could not revert. Please try again. / Could not redo. Please try again. / Action failed / File operation failed with "Couldn't roll back the files. Please try again." | The rollback or restore did not go through and nothing changed. Try again; if it keeps failing, close the drawer and open it again to refresh the list. |
| Download failed with Diagram is not ready to download yet | The diagram is still drawing. Wait for it to finish, then download it again. |
There are two common reasons. If the message box shows that someone is processing a request, another person is sending in the chat; wait for their answer to finish. If the box is greyed out, you lack Add Chat/Message, or Add Message In Others Chat in someone else's chat. Hover over the box to see the permission name, then ask your administrator for it. See Who can do what with Minds.
PII detection found personal data in the text and your company's policy blocks it. Remove or mask the data, then send again. When the policy only warns, Send it anyway appears for roles that may bypass the check. See PII checks in chat.
Check that Excel and CSV uploads are on for the Mind, that the file is 15 MB or smaller, that the conversation does not already hold 5 Excel or CSV files, and that the file is not a legacy .xls file. See Attach files and Excel data to a chat.
A Mind set to Execute as User, and every builder Mind, would run the SQL as the person you are viewing as, so DataMind OS refuses it. Exit the view to run queries. See Impersonate a user.
The task is already running, and a task cannot start twice at once. Wait for the current run to finish, then run it again. See Scheduled Tasks.
Click Retry to ask again. If the notice says Retry not started, read its description and try again. If a restart interrupted the answer, the chat offers Retry and Continue on the card. See Chat states and recovery and Ask a Mind.