Changelog

Organization and member settings are easier to manage

Organization and member settings now live in dedicated settings pages instead of being tucked into the command menu. That makes admin tasks easier to find, easier to revisit, and more in line with the rest of the product instead of feeling like secondary actions.

For teams that manage access often, this should feel like a meaningful cleanup. Settings have a clearer home, member administration is less buried, and common account-management workflows are easier to understand at a glance.

Large data sources are much easier to browse

The Data page is now dramatically more usable on very large connections. Instead of trying to load everything at once, schema and table browsing is much more incremental, which makes big databases feel far less likely to freeze or bog down the page.

Search and navigation are more intentional too. Browsing through lots of schemas should feel smoother, expanding sections is lighter-weight, and the command menu follows the same direction so exploring a large connection feels faster instead of overwhelming.

New dashboards can start with the right default access

Admins can now choose the default access pattern for newly created dashboards at the organization level. New dashboards can default to being available to everyone, only the creator, or the creator’s groups, with control over whether that shared access starts at manage, edit, or view.

This is especially useful for larger teams and embedded setups where “everyone can manage” is too open as a default. It gives teams a better starting point for governance without adding extra sharing cleanup every time a dashboard is created.

Automations can now be turned on and off cleanly

Automations now have a real active switch, so you can pause an automation without deleting it or rebuilding it later. When an automation is inactive, it stops scheduling runs, stops reacting to data-change triggers, and stops sending notifications.

That makes it much easier to temporarily pause an automation while you are testing, investigating, or waiting on upstream data. It also prevents half-disabled workflows where something looks paused in the UI but still keeps doing work in the background.

Fixes and improvements

• Simplified the charts page filters and included charts created from chat in the charts list.
• Improved time-series x-axis labels when SQL does not explicitly provide interval metadata.
• Fixed pie chart legend values so percentages and currency use the right formatting.
• Improved the SQL editor actions bar so it no longer obscures the bottom of the editor.
• Improved CSV exports for Excel on Windows so UTF-8 downloads open more cleanly.

Dashboards and insights feel much faster and more live

We reworked how dashboard charts and Insights load so the app does less unnecessary client-side work and updates more intentionally. Large dashboards should feel noticeably snappier, especially while editing, dragging charts around, or switching between tabs.

Insights benefit from the same shift. Loading is smoother, updates arrive more cleanly, and new content feels more immediate. We also tightened chart actions so duplicating or deleting a chart reflects right away instead of feeling delayed.

Number charts can compare two values

Number charts can now display a primary value alongside a comparison value. That makes them much more useful for common KPI patterns like current versus previous period, actual versus target, or value versus benchmark.

You can choose both fields directly in chart settings instead of working around the limitation in SQL or another chart type. The result is a simple KPI card that carries a lot more context at a glance.

Groups are now available through the public API

You can now list, create, update, and delete groups through the public API, and add organization members to them as well. This makes it easier to manage access-related workflows programmatically if you are syncing users and permissions from your own systems.

For teams building on top of Basedash, this closes an important gap in API coverage. Group management can now be automated instead of handled manually in the UI.

Fixes and improvements

• Added export support for chart drill-down tables.
• Improved embedded dashboards by fixing copy actions in embedded views and hiding copy-link actions that do not make sense there.
• Fixed SSO login issues that could block some SAML and embedded JWT flows.
• Fixed the demo data source onboarding flow so new connections load reliably right after creation.
• Fixed pie charts that could render cut off.
• Improved time-based x-axis labels so month and year boundaries read more naturally.
• Fixed connector logos so more data sources show the right icon.
• Fixed a crash on the new organization flow during full page loads.
• Improved the sidebar and dashboard context menu experience on large workspaces.
• Kept MCP-created chats out of the standard chats list so chat history stays cleaner.
• Fixed table picker row sizing so table selection feels less cramped.
• Fixed number charts so they still choose a sensible default value when one is not explicitly set.

Automations are much more capable

Automations now feel more like a first-class part of Basedash. The product language is now unified around automations, setup is cleaner, automations show up in command menu search, and you can give each one its own icon and color so important workflows are easier to spot at a glance.

We also added data-change triggers, so automations can run when a query result changes instead of only on a schedule. Sharing and permissions are more intentional now too, with clearer view and edit access for teammates, and automation pages refresh more smoothly while you are watching new runs come in.

AI can now follow your personal context

You can now add personal AI context from your profile, which gives Basedash guidance that is specific to how you work. This is a useful place to tell AI what metrics you care about most, how you want answers framed, or what business context matters for your role.

This personal context works alongside your team’s shared AI context instead of replacing it. That means AI can stay grounded in company-wide context while still adapting responses to each teammate across chat, chart creation, and other AI-assisted workflows.

Basedash feels noticeably faster

We cut a large amount of client-side weight out of the app and pushed more heavy code to load only when it is actually needed. That means less JavaScript up front, faster page readiness, and a lighter feel across common workflows.

The sidebar also renders much sooner now. Static navigation no longer waits on billing, sync, or live collaboration dependencies before appearing, so the app should feel more responsive instead of showing a blank sidebar while the page catches up.

Fixes and improvements

• Added a billing page that shows AI usage, included credits, and overages for the current billing period.
• Improved Slack AI conversations with clearer streamed progress updates.
• Improved chart drill-down reliability by validating drill-down SQL as charts are created and updated.
• Fixed automation trigger edits so saved data-change settings no longer get cleared unexpectedly.
• Fixed dashboard controls so values no longer carry across organizations.
• Fixed chart x-axes so small numeric values are less likely to be interpreted as dates.
• Simplified navigation by removing the standalone Charts sidebar item.

Members can now be deactivated and reactivated

Admins can now deactivate a member without permanently deleting their account history. This gives you a cleaner way to temporarily remove access when someone changes roles, goes on leave, or no longer needs the product day to day.

You can also reactivate that member later, which makes this much more practical than a delete-and-recreate flow. Existing ownership and linked data stay intact, so team administration is less disruptive when access changes over time.

Slack can now use approved tools

Slack conversations can now use MCP tools, which means Basedash can do more than just answer questions in Slack. When a tool needs approval, the approval flow happens right in Slack so you can keep moving without jumping back into the app.

We also extended tool access to background automations for tools that are explicitly allowed to run automatically. Taken together, this makes AI workflows feel more useful across both real-time Slack conversations and scheduled work.

Dashboards are easier to organize

Dashboards and folders are now easier to manage from the sidebar. You can quickly create folders from the dashboards page, reorder dashboards with a list-based drag-and-drop layout, and see cleaner folder nesting that makes larger workspaces easier to scan.

We also reduced some of the friction around navigation and sharing. Favorited dashboards still stay visible in the main dashboards list, and share buttons now show the current visibility state at a glance before you even open the menu.

Fixes and improvements

• Improved connection and table browsing with clearer breadcrumbs, easier settings access, and status dots for SQL databases and MCP servers.
• Improved chat and automation reliability so responses are less likely to stay stuck on “Generating…” when many automation runs are open.
• Fixed PDF exports for funnel charts and newer Tailwind color styles.
• Fixed DuckDB and S3 date handling so charts and AI see readable timestamps and dates.
• Fixed DuckDB PRAGMA table_info(...) and SHOW ALL TABLES queries when they include trailing semicolons.
• Fixed Data page record counts for DuckDB and S3 tables.
• Improved reliability for databases with strict connection limits by reducing default connection pool pressure.
• Fixed multi-select chart variable defaults so saved values behave more consistently.
• Added Redshift support for SSH connections.

Automations got a full refresh

Automations now have a cleaner, more structured home in Basedash. The redesigned experience adds a dedicated sidebar, a more useful header, and a command menu-driven setup flow that makes configuring automations feel more consistent with the rest of the product.

We also made automations easier to organize day to day. You can now favorite automations to keep the most important ones close at hand, and we fixed automation schedule editing so typing times feels natural instead of fiddly.

Chat is more reliable, more private, and better grounded

Ask mode is now better at understanding what you are looking at. When you open chat from a dashboard or a table, it can use that route context to stay grounded in the page you are already exploring instead of starting from a blank slate.

We also tightened up the rough edges that could make chat feel unreliable. This release reduces cases where chats stay stuck on “Generating…” after reconnects, prevents final answers from getting buried inside the steps accordion, and makes new chats private by default. Uploaded files and images are also handled more consistently, so exploratory workflows feel safer and smoother.

Browsing large data sources feels faster

Browsing schemas and tables is now much smoother on large data sources. The table picker is now virtualized, which keeps the UI responsive instead of rendering massive trees all at once as you explore bigger databases.

We also polished the browsing experience itself with cleaner hover states, easier copy actions, and improved dashboard and folder context menus. Altogether, moving around complex data sources should feel faster and less cluttered.

Fixes and improvements

• Fixed a line chart edge case that could make long category series loop backwards.
• Added a logout option during onboarding.
• Improved desktop app redirects.
• Prevented confusing dashboard filter edits from the chart editor when the filter belongs to a dashboard.
• Improved AI file uploads by normalizing file extension casing and fixing uploaded image rendering.

AWS Athena is now supported

You can now connect AWS Athena as a data source and start exploring and querying it like your other databases. This opens up a straightforward path to analyze data sitting in S3-backed lakes without needing to move it into a traditional warehouse first.

We also tightened up Athena-specific details (like the connection logo and SQL syntax edge cases) so the experience feels native end-to-end.

Schema Explorer is dramatically faster on large databases

Schema browsing and search are now significantly more responsive for big connections. The Schema Explorer UI is now virtualized, which eliminates the UI jank that could happen when expanding large schema trees.

Search is also more performant and intentional: very short searches no longer trigger expensive work, and searches cancel more cleanly while you type—so results feel snappy instead of laggy.

Embedded dashboards feel more complete

Full-app embed mode now supports the command menu (including the flows that rely on it, like filtering), but only shows the items that make sense in an embedded context. This fixes cases where embedded users would hit dead ends or “nothing happens” behavior because certain UI paths were blocked.

We also improved embedded polish: the loading screen now respects light theme correctly, embedded table charts scroll more naturally, and the embed sidebar behavior is more configurable (including optional hiding of Insights).

Insights are easier to tailor and act on

Insights can now have their own dedicated context, so you can guide how Insights should interpret your business and your data—without affecting other AI experiences.

You can also do more with Insights outputs: insight charts can be added to dashboards, and you can chat directly with insights to iterate and explore without losing the thread.

Fixes and improvements

• Enabled uploading CSVs and additional file types for AI workflows.
• Fixed alert schedule time entry so typing multi-digit hours works as expected.
• Prevented accidental overwrites when alert schedule values update while you’re editing.
• Fixed sharing menu search so results behave correctly.
• Fixed cases where the command menu could show “Data source not found” while a data source is still syncing.
• Improved reliability of chat/chart generation recovery after interruptions so messages don’t get stuck mid-generation.
• Fixed formatting and evaluation edge cases in SQL formatting.
• Fixed an Athena SQL offset syntax edge case.

Faster, more reliable live updates (Chat + Charts)

We made a substantial set of improvements to how Basedash streams live updates for chats and charts. The result is noticeably smoother real-time behavior—especially during longer sessions—reducing cases where updates stall, disconnect, or require a refresh to catch up.

These changes also improve the experience during deploys and intermittent network hiccups, so live content (like chat responses and chart updates) keeps flowing more consistently.

Infinite scroll for tables (no more paging)

Tables now support infinite scrolling, replacing traditional pagination. As you scroll, you can load more rows seamlessly—making it much easier to explore large result sets without jumping page-to-page.

This update applies across the product wherever tables appear (including dashboards and chart pages), and keeps the experience focused on exploration rather than navigation.

More control over Insights (including via API)

You can now disable Insights at the organization level—useful for teams who want tighter control over what appears in the UI and what gets generated. This setting is also configurable via the public API for teams managing orgs programmatically.

Fixes and improvements

• Added the ability to reorder filters.
• Improved variable creation/editing with AI-assisted enhancements.
• Reduced confusion during long-running data syncs by hiding certain “still syncing” connectors after 24 hours (purely visual).
• Improved the stability of the chart agent’s streaming responses.
• Improved chat message rendering resilience during partial updates.
• Strengthened security on authentication pages to better prevent clickjacking.
• Removed legacy “insight of the day” automations.

Control access to each data source (admin-managed)

You can now restrict who can query specific data sources inside an organization. Admins can grant access to “everyone”, specific groups, or specific people—while keeping existing orgs fully backwards-compatible (if you don’t configure grants, nothing changes).

This makes it much easier to onboard teammates safely (especially in larger orgs) without needing to split work across separate workspaces or duplicate connections.

Sharing & embed experiences are cleaner and more intuitive

Dashboard sharing permissions were streamlined into a clearer, single permissions list (instead of an “Everyone/Restricted” toggle), with improved behavior when multiple permission types overlap.

And for teams running Basedash in full-app embed mode, share menus are now hidden throughout the app—keeping embedded experiences focused and reducing “wait, should I click that?” confusion.

Fixes and improvements

• Added data-source access filtering across the product (so you only see/query what you have access to).
• Added an empty state in the chat sidebar when no chats match your current filter.
• Added an option to copy a chart’s SQL.
• Displayed Stripe discounts on the upgrade page (when a discount applies).
• Improved dashboard grid responsiveness.
• Fixed the filter popover opening automatically.
• Fixed the “Reset” button behavior on table views.
• Prevented empty SQL queries from running.
• Fixed “Copy as PNG” exports clipping dashboard metrics.
• Fixed metric layout overflow in some views.
• Improved sharing dialog layout to prevent overflow in tight widths.
• Added an SMTP “From email” field to self-hosted system settings (so admins don’t have to rely on an env var alone).

Refresh and organize charts faster

You can now refresh data for a single chart directly from the chart menu—handy when you’re iterating on one view and don’t want to refresh an entire dashboard.

We also made chart renaming feel much more lightweight: click the chart title in the header to rename it right in place, so keeping dashboards tidy is a lot quicker.

Stay in flow while AI works

Basedash can now send desktop notifications when AI finishes, so you can switch tabs (or step away) and still know exactly when results are ready.

We also added clearer loading states while chat messages are loading, making conversations feel more responsive—especially on slower connections or when opening longer threads.

Fixes and improvements

• Fixed dashboards briefly showing older cached data after a refresh/reload.
• Improved how AI formats percentage values so results read correctly as 0–100% (instead of 0–1).
• Improved AI SQL generation across dialects with clearer dialect-specific guidance.
• Improved reliability of AI query execution by automatically correcting certain tool-call errors.
• Improved stability of live streaming updates for chats and chart runs (long-lived connections now recover more gracefully).
• Reduced duplicate/overlapping data sync runs (including debouncing repeated manual sync clicks).
• Made periodic data syncing more consistent by moving it fully server-side.
• Improved automation reliability and scalability so automations are less likely to miss runs under load.
• Fixed a background job edge case that could cause intermittent failures.