Dashboards

A dashboard is a maintained output built by an agent. Use it when people should return to the current result and its sources.

Dashboards can use widget boards or apps. If your dashboard shows Draft, editors review app changes before they replace the current view. Dashboard apps are beta. Widget boards keep their widgets, layout, and board history.

Creating a dashboard starts its first build immediately. Enter a name and instructions, then choose an optional project, model, and available reasoning level. Describe the audience, sources, questions, and presentation you need. If you choose no project, the dashboard keeps its own context. A catalog model choice follows its newest eligible family version on later work. Running work and history keep the version used.

Review and publish an app dashboard

You need dashboard edit permission to review drafts, publish changes, or restore a version. Viewers can open the published dashboard and saved history. Drafts and publication controls are available only to editors.

1. Open the dashboard and wait for Building the first version to finish.
2. Select Draft and test its controls and data.
3. Ask for corrections in the dashboard chat if needed.
4. On the web, select Make dashboard available when the first draft is ready. On mobile, select Publish dashboard.

The current version is what viewers open. Later agent changes prepare another Draft. Review it, then select Update dashboard on the web, or Publish changes on mobile. Finishing an agent response does not publish an app. Publishing does not make the dashboard public or change who can access it.

The web app header keeps the view switch and the open-in-new-tab icon together. History is in the dashboard's options menu. Opening a new tab keeps the same access requirements; it does not create a public link.

The new tab shows the app under a slim d5s bar. Select the d5s mark to go to the workspace, or the dashboard name to switch to another app dashboard, copy the link, or open the full dashboard page. Editors see Chat, which opens the dashboard conversation. People who can manage the project's access see Share, which shares the whole project. A saved version shows its version number, and a draft shows Draft.

Use Chat, Split, or Dashboard to move between the conversation and the app. On a phone, use Chat or Dashboard. Switching views keeps the open app while it remains available. Save important input through the app's own controls before leaving or reloading.

Type @ in dashboard chat and select an accessible workspace agent. Its reply and changes stay in the shared dashboard conversation under its name. Its private conversation and credentials are not shared. You still review and publish the resulting draft. In the web chat panel, Contributions opens the selected contribution.

Ask the agent in dashboard chat to test the ready draft in its browser before you publish. An agent opening the dashboard from another conversation sees the published app when one exists. Testing a draft does not publish it.

If you see Draft needs a fix, select Ask agent to fix. An existing published version stays available. A failed publication also keeps the previous published version. Review the error and draft before trying publication again.

Update an app's data

Ask an agent in dashboard chat to update the data or change how the app gets it. Review any new draft before choosing Update dashboard.

An app can also read updated saved data and its permitted live sources without a new app version. It must request the data again to show changes. Reload the app if its display does not update. Source access and supported read operations still apply.

Preview and restore an app version

Select History. On the web, it is in the dashboard's options menu. Select Preview beside a saved version. Historical previews are read-only. They show captured data and cannot fetch live source results. Select Back to current on the web, or Back to published on mobile, to return to the current version.

To restore, preview the version, open version history again, and select Restore this version. Restore creates a new published version without asking the agent to rebuild it. The published version can use refreshed data again. Current source permissions still apply. If restore fails, the existing publication stays in place. If someone has published another version meanwhile, inspect that version before retrying.

Reopen a stopped app dashboard

If the page says Your dashboard is stopped, select Start dashboard when available. Wait for Starting your dashboard to finish. This opens the selected saved version; it does not ask the agent to rebuild or publish it.

For Dashboard is unavailable, select Try again. If the error continues, share its visible message in dashboard chat. For This saved version is unavailable, return to the current version. If access has changed, ask the owner to review your permission. An open app does not retain access after that permission is removed.

Work with an existing widget board

The remaining instructions apply to dashboards made from named widgets. Their board publication and history controls remain available.

Refreshable material

Dashboard instructions describe the audience, data to use, questions to answer, and presentation the agent should produce. The agent uses native metric, chart, table, and text widgets for structured results, and custom HTML widgets for interactive views. Each widget has a stable name and can be updated without replacing the rest of the board.

Use Chat, Split, or Dashboard to switch between the conversation and board. Ask the agent to update a named widget in the existing conversation. Select Sources in the dashboard's options menu to refresh saved connector datasets independently of the design. Select Arrange in the same menu to move widgets, then select Done arranging. Completed agent responses use the same personal feedback controls as chats, including an optional details dialog. The chat panel shows only your own selection, not teammate ratings or aggregate counts. Each published board update records a new board revision.

A manual refresh updates the current board from its sources. It does not rebuild the board from the original request. The agent refreshes existing source-backed datasets and changes the board only when fresh data or a new explicit instruction requires it.

When messages wait behind an active dashboard build, the chat panel shows them inside the composer, three at a time. Choose N more to inspect more, Steer to bring a waiting message forward and ask the active build to stop, Edit to put a waiting message back in the composer, or × to delete that waiting message.

The dashboard uses the same reply composer as ordinary chats. Attach files with +, paste, or drag and drop. Use the controls available to you in the composer. Failed sends keep your draft and attachments. Retrying an unconfirmed message does not create a duplicate.

Ask and Allow all control connector actions in this dashboard conversation. The choice also applies to later work in that conversation. Switching to Allow all retries a waiting approval once. Account permissions, explicit denials, and organization restrictions still apply. Email sends and calendar bookings remain unavailable in dashboard work. See chat approval controls for other action limits.

The dashboard chat panel also supports /compact for an existing idle build session. Choose it from the compact / menu, then submit again to summarize eligible older conversation context without requesting a dashboard update or creating a chat bubble.

Registered API data sources

Dashboard data sources use supported read operations from a connection available to the dashboard's project. Current connection permissions apply to every refresh. Widgets cannot request arbitrary external destinations or provide credentials.

GitHub pull-request lists are supported for direct API data sources. Connecting an account does not make every API operation available as a dashboard source. MCP sources also require a supported read operation and structured data.

GitHub data sources

A GitHub data source can list open pull requests for a repository using a GitHub connection available to the dashboard's project. Choose the summary fields to publish, such as pull-request number, title, author, and update time.

Refreshing a data source updates its saved data without starting another agent conversation. If access is revoked, the provider fails, or a complete result cannot be collected, the previous complete result is retained. A data source holds at most 1,000 rows. If a source goes over a limit, Sources shows the limit that stopped the refresh.

Let an agent add dashboard data

Ask an agent to add records to a dashboard, such as findings, customer updates, or progress notes. Identify the dashboard and the fields to keep. The agent needs permission to edit it.

For example: “Add this finding to Security assessments. Keep its source, observation time and stable finding ID. Update that finding if it already exists.”

An app must first be connected to those saved records. Once connected, it can show data updates without a new app release. Refresh the view to load the latest records. An app that only displays an imported file needs a one-time update to use saved records.

Use the app’s data-entry form for people and browser agents. Ask its builder to add the form and enable writes for the connected record dataset. This needs a new app draft once. Saving records then updates the existing dataset without rebuilding the app. Only dashboard editors can save through a current preview; saved historical versions remain read-only. Saves from draft and published forms update the shared records immediately. Use separate Demo data when testing a form. An agent opens accessible dashboard links through its assigned browser using its current task permissions. It does not need a separate human sign-in. Browser access must be enabled. View-only access cannot save records; stopping the task, taking private control, or revoking access ends that agent preview. The agent verifies a save by refreshing the view. Direct record tools remain available when direct entry is requested.

If your agent shows Computer, ask it to find an accessible dashboard and open it in its browser. It can inspect the page and use its visible controls. Your task and dashboard permissions still apply. You do not need to share a human sign-in.

Repeating an identical import does not duplicate records. If another edit changes the data, the agent must inspect it before applying a conflicting update. Keep Demo records separate from real workspace data. Saved versions retain their captured data.

Record datasets support up to 1,000 records and 1 MB of data. Each record has declared text, number or boolean fields. Records are visible to dashboard readers, so do not include secrets.

Publishing an HTML widget

Ask the dashboard agent to publish a saved HTML report as a named widget. The widget appears once its HTML and supporting files have been captured successfully. An existing widget stays visible while its replacement is being prepared.

If the board changes during publication, or the files cannot be captured, the previous widget stays in place. The agent can inspect the current board and submit a new publication when you are ready to try again.

When someone views the board, d5s records the first runtime error of each HTML widget. The agent sees this error when it inspects the board, and it can publish a fix.

Custom HTML views

A custom HTML widget can use the same saved datasets as native charts and tables. Refreshing a source updates those views without asking the agent to redraw the dashboard. Each source keeps its own refresh state and last successful timestamp.

Leaving the saved report stops live updates.

HTML widgets can follow the app theme and automatically fit their content height, within a bounded preview. Longer reports scroll inside the preview. A frameless presentation lets a report use its own heading and layout and starts at the full board width. Updating a widget preserves your saved arrangement. Narrow boards, including Split view, stack widgets so each remains readable. Source links appear as an Open source link outside the embedded report, where you can open the provider page.

Rename a dashboard

Choose Rename from a dashboard's actions menu, on the Dashboards list, a project's Dashboards strip, or the dashboard page itself. The title stays editable while a build is running and after one fails.

Renaming updates the title wherever the dashboard and its build sessions appear. A project you named stays unchanged.

Versions keep history

Select History in the dashboard's options menu to see published board snapshots, including widget content and layout. Preview a snapshot to compare it, or restore it without asking the agent to rebuild the board. Historical HTML views use captured files and do not connect to live datasets.

Reports, CSV datasets, and source code presented in the conversation remain attachments. Opening an attachment shows a separate, closable preview; it never replaces the board. Only an explicit named-widget publication changes dashboard content. A completed agent turn without published widgets is not a live dashboard.

The project supplies reusable instructions, memory, files, skills, connectors, and access. The dashboard owns its instructions, conversation, widgets, saved datasets, layout, and board history. Deleting a dashboard preserves its past build sessions.

Other workspace runs can discover and inspect a dashboard, then contribute by updating its title, description, instructions, or enabled state. Publishing widget content remains an explicit operation in the dashboard conversation.

Native reports

Ask the agent to group an explanation, headline metrics, charts, and supporting tables into one report widget. Reports use the app's own components and move together on the board. Their values are saved snapshots: ask the agent to update the report when its analysis changes. For independently refreshed data, use a dataset-connected metric, chart, or table widget.

Interactive read panels

Custom panels can offer named reads from a saved dataset or a live connector. For example, select an issue status and choose whether to read the saved snapshot or fetch current results. These reads run without an agent rebuilding the panel and do not replace saved dataset snapshots.

The panel can use only its configured sources and permitted filter choices. If the source definition or access changes, the read stops until the panel is updated. Live read permissions must be granted by the source owner, including when restoring an older panel from history.

Work with a workspace agent

Type @ in dashboard chat and select an accessible workspace agent. Its reply stays in the dashboard conversation. Include a widget selection when the request concerns that widget. Its private conversation is not shared. The agent can inspect accessible widgets and datasets as needed.

You can also type @ in the dashboard composer to insert an inline reference to a colleague or accessible project, conversation, or dashboard. Choose an agent in the same @ menu to request its reply. A colleague reference does not grant dashboard access or send a chat-mention notification.

Open Contributions to read a contribution and see its responding agent. When a named contribution ends, the agent’s own conversation receives a compact dashboard card. For people who can still open the dashboard, the card shows the dashboard name, agent, time, the verified outcome, and any widgets known to have updated or failed to publish. A pending HTML publication stays marked as publishing until its result is known. Without dashboard access, the conversation keeps only a generic activity entry and link. The request and contribution transcript are never copied into the agent’s conversation.

A follow-up starts with current dashboard context. Files from a contribution remain available there.

You can also ask an agent from its own chat to inspect and update a dashboard you identify. The original chat stays in its project, and dashboard access is checked separately. If someone changes the same widget while an edit is underway, review its current version before retrying.

Workspace membership does not grant another agent's credentials. Current dashboard, conversation, and agent permissions still apply, as do the selected agent's budget and model policies.

MCP data shapes and limits

MCP data sources can use supported row lists, arrays, or single records. A result may cover only part of the remote account; confirm the source's scope before treating a dashboard as complete.

Refresh requires an allowed, read-only operation. Prose, images, files, and mixed content can inform an agent's analysis but are not automatically refreshable datasets. Verify the first refresh and the displayed data before relying on a live widget.

Google workspace protection

Google content can be used for the work you authorize in this workspace. Tasks using AI may send relevant content to the AI providers used by the workspace. Disconnecting does not erase information already retrieved. See Google data protection for AI processing and stored-data guidance.