🐛 Page previews from a partial load their queries and filters
Using View another page while editing a partial now loads every query on that page, so filters, dropdown options and conditionals work as they do on the page itself, and the partial you’re editing runs with the values the page passes in throughvariables=.✨ Slack bot answers without an @mention in chosen channels
Admins can pick Slack channels where the Evidence bot answers data questions without being @mentioned.- Add channels in Settings → Integrations → Slack, searching among the channels the bot has been added to.
- The bot only answers messages that ask for data, and skips anything that tags a person.
✨ French chart date labels
Chart axis date labels, tooltips, table date columns and customfmt date formats now render in French when a viewer selects fr from the translations menu.- Date-grain defaults follow French order and a 24-hour clock (
15 janv. 2025,15 janv. 2025 16 h,T1). - Fixes an English axis label that showed
Janinstead of the tick’s own month when a chart’s first visible tick of a new year was not in January.
✨ Built-in report text in French
Tables, filters, date pickers, legends and other built-in component text now render in French when a viewer selectsfr from the translations menu.- Covers table pagination and search, filter chips and operators, dropdown and date range picker labels and presets, date grain and comparison selectors, legend controls, download menus, sampling notices, commentary timestamps, calendar heatmaps, map labels and computed “Total” labels.
- Table comparison column headers use French finance conventions (
vs N-1,vs M-1,Objectif), including in scheduled PDFs, and comparison tooltips are translated too. - Reporting periods read in French too (
juil. 2026,T2 2026,Semaine du 2 août 2026): in the period picker, at the top of scheduled PDFs, in{{period}}text, and incustom_rangesnames. URLs anddefault_rangevalues are unchanged. - Attributes you set explicitly (for example
total_label) keep your text in every language. - In every language, the date range picker calendar now starts weeks on the project’s first day of week, and its preset list widens to fit longer labels.
- Adding another language is a small open-source contribution: see the translations docs.
✨ Amazon Athena direct connector
Query Amazon Athena directly from Evidence, without syncing data first.- Connect with an AWS access key, with an optional IAM role to assume and a workgroup, catalog, and database.
- Row-level security: assign each Evidence user an IAM role under Settings → Access Rules, and your Lake Formation filters decide what they see.
- Athena query result reuse is on by default, so repeated queries skip the scan.
- Works in the CLI:
evidence init --warehouse athenascaffolds aconnection.yaml, and local queries run against Athena using your AWS profile or keys.
🚀 AI overage budget
Admins can now cap AI usage-based pricing with a monthly overage budget in dollars.- In AI settings, with usage-based pricing enabled, set the most you’ll be billed beyond your included credits each month.
- The settings page shows how many additional credits the budget allows beyond your included credits.
- Once the budget is reached, AI pauses until the next billing period or until an admin raises the budget.
- Leave the budget blank to keep usage-based pricing uncapped, as before.
✨ Open external links from HTML blocks and modals
Links in{% html %} blocks and their modals now open in a new tab, and scripts can call evidence.openUrl(url) to send a click on a card or map marker straight to an external tool.<a href="https://…">opens in a new tab;target="_blank"works too but is no longer required.<a href="/project/page">drills through to another page in the app, the same as a markdown link.evidence.openUrl(url)opens an absolutehttp(s)URL in a new tab from a click handler. The report page itself is never redirected.- A block can only open hosts it names. Any
https://hostwritten in the block is allowed automatically; when the URL comes from data, list the host withlinkHosts="support.example.com"on the tag.
🚀 Pick individual repeat sections for PDFs
The PDF contents picker now lists exactly the headings that will print, including a heading for eachrepeat value, so you can include or exclude them one by one.- Sections inside
hidden_in_printare no longer listed; sections insideonly_in_printare. - Headings with
{{ }}variables show their resolved text, andif/elseblocks list only the branch that will render. - A heading inside a
repeatis listed once per value and can be unticked on its own; new repeat values are included by default in scheduled reports. - The “On this page” table of contents lists each repeated heading with its resolved text.
✨ SQL Server direct connector
Query SQL Server, Azure SQL Database, and Azure SQL Managed Instance directly from Evidence, without syncing data first.- Connect with a SQL login: server, port, database, user, and password.
- TLS is on by default, and you can supply your own CA certificate for servers with a self-signed or internal-CA certificate.
- Row-level security: your SQL Server security policies can filter data per Evidence viewer.
- Works in the CLI:
evidence init --warehouse sqlserverscaffolds aconnection.yaml, and local queries run directly against your server.
✨ Preview pages at mobile width in the editor
A new toggle on the editor’s preview pane narrows the preview to phone width, so you can check how a page reads on mobile without resizing your browser.- Mobile-specific behavior shows in the preview too:
detailshonorsopen_mobile, and modals open as they would on a phone, inside the preview. dimension_gridnow wraps based on its own width rather than the browser’s, so it lays out correctly in narrow columns as well as on phones.
✨ Commentary per input selection
Commentary can now be kept separately for each selected value of one or more inputs, alongside the reporting period.View Documentation✨ Period selector component
Added a period selector input, which can also place a page’s reporting period anywhere on the page.View Documentation🚀 Improved row layout
Rows can now spread their contents across the full width, for example a heading on the left and a button on the right.View Documentation✨ Generate Excel Report
A new Generate Excel Report option exports the page as an Excel workbook with a snapshot of the report on the first sheet and a data sheet per component; Download Data now also appears in the CLI.- Download Data is unchanged and still exports data only; pick Generate Excel Report when you want the report snapshot included.
- Hide it with
downloads: { excel: false }in page frontmatter. It is also hidden whenimageordatadownloads are turned off. - The local CLI viewer now has Download Data and Generate Excel Report in its page menu.
🚀 Set down_is_good for a whole page
Adddown_is_good: true to a page’s frontmatter to treat decreases as good for every delta on that page.- Applies to
big_value,delta, and table delta columns. A component’s owndown_is_goodstill overrides it. down_is_good=trueon abenchmark_comparisonortarget_comparisonoption is now respected when a component uses thatcomparison_selector.
🐛 Tables inside repeat blocks now read page inputs
Atable inside a {% repeat %} block now resolves page inputs referenced in its where, having, order, and qualify attributes instead of running with the value missing.- Previously the table only looked at the repeat’s own scope, so
{{my_input}}in these attributes interpolated as empty and the query failed with a syntax error. - Titles and other text attributes on the same table also resolve page inputs correctly now.
🐛 CLI runs direct-connector projects without reaching Evidence Studio
evidence dev no longer refuses to start when evidence.studio is unreachable for projects with a connection.yaml, evidence upgrade now works on Windows, and the home page honours page_width and cards like every other page.- Projects with a
connection.yamlskip the Studio reachability check entirely; projects on the Evidence Warehouse get a warning (with the failure reason and a proxy hint) instead of a hard exit. evidence devprints which mode it is in — direct connector or managed warehouse — so it is clear whether a login is needed.evidence upgradesupports Windows x64 and Linux arm64, keepsevd.exein sync on Windows, and points to the installer when the update check fails.evidence serveis listed inevidence help.pages/home.mdnow respectspage_width,cards,table_of_contentsand frontmattertheme, matching subpages.
🚀 Details start state on mobile
Thedetails component has a new open_mobile attribute to choose whether it starts open or collapsed on mobile screens, independently of openView Documentation✨ Validation summary in the editor
A new indicator in the editor toolbar shows how many validation errors are on the open file. Click it to see every error with its line number, and click an error to jump to it in the editor.- Covers component errors, inline SQL query errors, and config file (yaml) validation.
- Each entry is labeled with the component, query, or config key it belongs to.
- Clicking an error scrolls the editor to the line and highlights it.
✨ Commentary supports workflow periods
Commentary components can now keep separate notes for each reporting period withscope="period".- Period-scoped commentary follows the page’s
workflow.periodpicker, including in PDF renders. - Commentary without
scoperemains page-wide.
✨ Waterfall chart
Thewaterfall_chart component is now available - useful for growth decomposition, revenue bridges, and reconciliationsView Documentation🚀 Resume GitHub Enterprise connections
GitHub Enterprise Cloud connections can now resume after another organization owner installs the app, with clearer errors and controls for unfinished attempts.- Finish connecting from GitHub integration settings when the installation callback is missed.
- Allow up to seven days for installation after app registration.
- Resume or cancel an unfinished connection instead of creating another app on each retry.
- See an error message and request ID when a connection needs support.
✨ Reporting periods
Turn a page into a periodic report with aworkflow.period block in frontmatter.- Adds a period picker at the top right of the page, listing complete calendar periods (day, week, month, quarter or year)
- The selected period is available as the
periodvariable —{{ period.label }}in markdown,{{ period.between }}in SQL - Only finished periods are offered, so a published report doesn’t change while its period is still running
- The selection is kept in the URL, so a link points at a specific period
✨ More control over color scales
Color scales in tables and map layers now support diverging midpoints and custom value-to-color breakpoints, and behave consistently across both.- Set
midpointto anchor a diverging scale at a value such as 0, andmin/maxto fix the scale bounds. - Use
color_stopsto pin specific colors to specific values. - Diverging scales stay correct even when every value falls on one side of the midpoint.
✨ Scheduled Reports
Email reports to your users on a regular schedule, or send one-off reports on demand.View Documentation✨ Access Viewer AI inside Slack
Connect a Slack workspace in Integration settings and message the Evidence bot to get answers from your data without leaving Slack.- An admin adds the app from Settings → Integrations → Slack, then each person connects their own account — the bot only reads what you can already see.
- Direct message the bot, or @mention it in a channel.
- Answers can display charts, save insights to your library, and link back into Evidence.
- Admins can turn the bot off for the whole organization, and choose whether each person’s first answer in a channel is shown to them before the channel sees it.
🐛 Sorting a table by a grouped date column
Sorting a table by a date column that has adate_grain now sorts by the grouped date instead of failing with a grouping error.order="my_date desc"alongside{% dimension value="my_date" date_grain="day" /%}sorted by the raw column, which the warehouse never grouped. ClickHouse rejected the query and the table rendered empty.- A hidden sort column on a table with subtotals is now aggregated, so it no longer breaks the subtotal and total rows.
✨ Bring your own model API key
Enterprise organizations can now point both AI agents at their own OpenAI, Anthropic, AWS Bedrock, or Google Vertex account from Settings → AI Agent.- Both the editor agent and the viewer agent use your provider and model, so your provider bills you for inference directly and Evidence AI credits are neither consumed nor charged.
- Credentials are stored in the WorkOS Vault and never sent back to the browser. Saving verifies the credential against your provider first, so a wrong key or model ID is rejected rather than stored.
- There is no fallback to Evidence’s models: if your credential stops working, chat reports a clear error instead of quietly switching back and billing you for credits.
✨ Reports page for run history and schedules
A new top-level Reports page (next to Models) for viewing report run history and managing schedules.- History tab: View all report runs with status, trigger type (manual/schedule/API), author, timing, and recipient count
- Schedules tab: View and manage recurring report schedules with pause/resume and delete actions
- Admin-only actions: Pause/resume schedules, delete schedules, re-run past reports, download PDFs
- Developer access: Developers (members) can view the page; admins can take all actions
- Entitlement gate: Available on Team, Pro, and Enterprise plans
✨ GitHub Enterprise Cloud support
Enterprise organizations can now connect Evidence Studio and the Evidence CLI to GitHub Enterprise Cloud accounts hosted on dedicated*.ghe.com domains.- Create and install a private, organization-owned GitHub App through Organization Settings.
- Select accessible enterprise repositories in Project Settings and use Studio branches, pull requests, checks, and publishing as usual.
- Run
evidence launchorevidence linkfrom HTTPS and SSH GHE.com remotes.
✨ MotherDuck direct connector
Connect Evidence directly to a MotherDuck database to build reports on live data without syncing it first.View Documentation✨ Custom Maps
Use thecustom_map component to build fully customized maps with Mapbox GL or MapLibre GL.- Write map code directly in the component body — pull page data with
evidence.query, read and write page filters, and react to variables and theme changes. - No key needed by default: keyless MapLibre out of the box, or Mapbox using Evidence’s shared token under the hood.
- Import map-ecosystem plugins (mapbox-gl-draw, deck.gl, turf, and more) straight from a CDN.
🐛 Custom HTML and ECharts blocks work in the CLI
{% html %} and {% custom_echart %} blocks now render when running a project locally with the Evidence CLI.- Both blocks previously rendered blank in
evidence dev— the sandbox runtime they load was only built by the hosted app, so the CLI served nothing for it - Requires CLI v0.7.4 or later
- If a sandbox runtime is ever missing, the block now reports an explicit error instead of staying blank
✨ Cube Connector
Connect Evidence directly to your Cube semantic layer through its Postgres-wire SQL API.- Query your cubes and views live from Evidence — metrics stay defined in Cube.
- Configure it in Connectors → Warehouse → Cube (host, port, user, password, SSL), or via
connection.yaml(type: cube) in the CLI. - The SQL editor validates expressions against Cube’s documented function subset, so unsupported functions are flagged before you run a query.
✨ PDF “Use page colors” option
Preserve the exact colors from your page when generating PDFs.- New “Use page colors” checkbox in the PDF settings modal (Cmd+P)
- When enabled, PDFs retain the theme’s actual background colors (including dark mode or custom theme backgrounds)
- When disabled (default), PDFs use white backgrounds optimized for printing
- Preview updates to show how the PDF will look with the selected option
✨ Postgres Direct Connector
Connect Evidence directly to Postgres — including Amazon RDS/Aurora, Supabase, Neon, and Timescale — and query it live.- Add Postgres under Settings → Warehouse: enter host, port, database, user, password, and SSL mode, test the connection, and save.
- Works with any Postgres-wire source: generic Postgres, Amazon RDS / Aurora, Supabase, Neon, and Timescale.
- Full SSL/TLS support with
disable/require/verify-ca/verify-fullmodes, plus an optional CA bundle and mutual-TLS client certificate. - The schema browser lists tables, views, and materialized views.
- Also available from the CLI via
connection.yaml(type: postgres) orevidence init --warehouse postgres, with certificates referenced by file path.
✨ Evidence Studio MCP Server
Search your reports and query your data from your favorite AI client. Connect Claude, ChatGPT, and other MCP clients to your live Evidence workspace — every request runs as the signed-in user, so access policies and page controls are respected.View Documentation🐛 CLI dev server no longer crashes when a subfolder has an index.md
Nestedindex.md / home.md / README.md files are no longer treated as the home page in the sidebar, so evidence dev renders normally instead of crashing with a duplicate-key error when a project has a root home page plus an index.md in a subfolder.✨ Lemlist Connector
Connect your Lemlist data to Evidence. Sync campaigns, leads, and activities from your email outreach.View Documentation✨ Table Access Restrictions
Restrict tables to a chosen set of usersRestrict specific tables to a chosen set of users from the Access Rules settings page. Unrestricted tables stay open to everyone in the organization, and restricted tables show a lock badge in the editor’s schema list.View Documentation🐛 Y-axis titles stay legible above chart data
Chart contents no longer overlap the y-axis title — bars, areas, and lines that reach the top of the axis now render beneath the title instead of covering it. Gridlines are cut off cleanly behind the title rather than showing through the text, the title is vertically aligned with the top axis tick label, and a subtle text outline keeps it readable when data passes directly beneath it. Applies to both left and right (y2) axis titles.🐛 range_calendar “all time” no longer breaks raw SQL
Referencing{{ date_filter.between }} in a where or SQL template while “all time” was selected previously produced an empty string, which broke the query (where date → syntax error). .between now emits IS NOT NULL for “all time”, so where="date {{ date_filter.between }}" matches all rows and just works.The new all_time_range attribute controls what “all time” means for a specific range_calendar:all_time_range="none"restores the previous empty-string behavior — pair it with[[ ]]conditional blocks or a| fallbackwhen the whole clause should drop (e.g. to keep NULL-date rows, whichIS NOT NULLotherwise excludes).- Set it to a bounded range — a preset key (
"last 12 months"), acustom_rangeslabel, or a range expression ("from 2022-12-01") — and “all time” resolves to that range everywhere (.between,.start,.end, and thedate_rangevalue). If you were faking this with a| fallbackin yourwhereclause, setall_time_rangeto that range and delete the fallback.
.start / .end stay empty for unbounded “all time” (so date spines and axes fail loudly rather than silently wrong), and .filter / a filters=[...] reference stay a true no-op that keeps NULL rows.✨ Extra tooltip rows on chart series
Addtooltip_fields to any bar_chart / line_chart / area_chart / scatter_chart / bubble_chart — or to bar, line, area, scatter, bubble inside a combo_chart — to surface additional aggregated values on hover, with optional labels, format codes, and color_by_sign / down_is_good for delta-style coloring.✨ ClickHouse direct connector
Connect Evidence directly to your own ClickHouse instance to build reports on live data without syncing it first.View Documentation✨ Vibe code any visualization with the custom_echart component
Need a chart Evidence doesn’t have out of the box — a waterfall, a bridge, a radar, something bespoke? The newcustom_echart component lets you build any visualization you can imagine. Describe what you want to the assistant and it writes the chart for you, wired up to your data, filters, and theme automatically.🚀 HubSpot Connector
Updated the HubSpot connector with more reliable syncs and broader table coverage.View Documentation✨ Microsoft Fabric direct connector
Connect Evidence directly to a Microsoft Fabric Warehouse or Lakehouse SQL endpoint to build reports on live data without syncing it first.View Documentation✨ Snowflake and BigQuery direct connectors
Link Evidence to Snowflake or BigQuery directly to build reports without syncing data.View Documentation✨ Evidence CLI now available on Windows
The Evidence CLI now ships a Windows x64 binary, installable via PowerShell:irm https://gaamozau3jchzs3r.public.blob.vercel-storage.com/cli/install.ps1 | iex.View Documentation
✨ YouTube Analytics Connector
Connect your YouTube Analytics data to Evidence. Sync channel reports, content owner reports, and metadata tables.View Documentation✨ SAP HANA Connector
Connect your SAP HANA data to Evidence. Sync data from SAP ECC and SAP S/4HANA systems running on the HANA platform.View Documentation✨ Evidence CLI
Develop Evidence Studio projects locally from the command line withevidence dev.View Documentation
✨ PayPal Connector
Connect your PayPal data to Evidence. Sync transactions, invoices, payments, subscriptions, and billing plans.View Documentation
✨ RevenueCat Connector
Connect your RevenueCat subscription data to Evidence. Sync subscribers, transactions, entitlements, offerings, and products.View Documentation✨ Candlestick chart component
Added a new candlestick chart component for visualizing OHLC financial data with optional volume bars.
✨ Percentage labels for pie charts
Addedpct and pct_fmt options for pie charts to display percentage values directly on slice labels.
✨ Salesforce Connector
Connect your Salesforce CRM data to Evidence. Sync accounts, contacts, opportunities, leads, and campaigns.View Documentation✨ Gradient fill for area charts
Area charts now support a gradient fill option that fades from the series color at the top to transparent at the bottom. Enable it witharea_options={ gradient=true }.
🚀 Model Engine and Primary Key Settings
The model edit page now includes configurable refresh schedules, table engine selection, and primary key configuration with helpful tooltips explaining each engine type, data locations, and nullable primary key handling.View Documentation✨ Datadog Connector
Connect your Datadog monitoring and analytics data to Evidence. Sync metrics, dashboards, monitors, hosts, and usage data from your Datadog account.View Documentation✨ RDS PostgreSQL Connector
Connect your RDS PostgreSQL database to Evidence and sync your data automatically.View Documentation✨ Collapsible Tables
Tables with subtotals can now be collapsed. Click subtotal rows to expand or collapse their children.
✨ Total and Subtotal Position
Control where totals and subtotals appear withtotal_position and subtotal_position. Set to “top” or “bottom”.
🚀 Clear Table Sorting with Third Click
Click a column header a third time to clear sorting and return to the original order.View Documentation✨ Azure SQL Database Connector
Sync data from Azure SQL Database directly into Evidence.View Documentation✨ Logo Attribute for Table Dimensions
Display company logos in table dimension columns. The newlogo attribute accepts a column containing domains to look up logos, and logo_options lets you customize size, grayscale mode, and label visibility.
✨ Twilio Connector
Connect your Twilio data to Evidence. Sync calls, messages, accounts, and usage records.View Documentation🚀 BigQuery 1 Minute Incremental Syncs
BigQuery data connector now supports 1 minute incremental syncs for faster data freshness.View Documentation
✨ Google Ads & Facebook Ads Connectors
Connect your Google Ads and Facebook Ads data to Evidence. Sync campaigns, ad groups, ads, and performance metrics to build marketing analytics dashboards.View Documentation✨ PostHog Connector
Connect your PostHog product analytics data to Evidence. Sync events, persons, cohorts, feature flags, insights, and more from your PostHog account.View Documentation🚀 Snowflake 1 Minute Incremental Syncs
Snowflake data connector now supports 1 minute incremental syncs for faster data freshness.View Documentation✨ Smooth Line Option
Newsmooth attribute inside line options lets you create a line with rounded corners
View Documentation✨ Chart height attribute
Control the height of your charts using the height attributeView Documentation✨ Customer Groups
Create customers, and invite customer users to view reports that you choose.
View Documentation✨ User Variables
Access the user’s name, email and organization using variables in markdown.View Documentation✨ Themes
Set color themes at the organization, project, and page level - including background, chart, and card colorsAccess theme settings in page settings, project settings, or org settings. You can override theme colors at each level (e.g., project can override org theme, page overrides project and org themes)✨ Translation Management
Define variables with different translations for multi-language apps.
View Documentation✨ Standalone Horizontal Bar Chart Component
New dedicatedhorizontal_bar_chart component makes it easier to create horizontal bar chartsWe recommend migrating any existing horizontal bar charts to this new componentView Documentation✨ Display control for total and subtotals in table
Choose which totals or subtotals to display in your tableView Documentation
✨ Input tabs full_width and align attributes
Input tabs now allows the same styling configuration as the Tabs component, including full width and right alignView Documentation
✨ fmt property for input options
Use dynamic formats in viz components when creating an input to select a metric - e.g., show ‘usd’ for sales or ‘pct1’ for growthView Documentation✨ Variable label property
You can now reference the label property of a variable - useful when adding labels to viz components based on input selectionView Documentation
✨ Toggle input
New toggle input component which sets a boolean variable you can use in queries and components
✨ Button group and input tabs
New input components as visual alternatives to dropdowns:button_group and input_tabsNote that this update also deprecates the dropdown_option component in favour of the more general option component. The option component can be used inside dropdown, button_group, or input_tabs
✨ Stacked bar series in combo charts
Bar series in combo charts can now be stacked using the newstack_id attribute. Bars with the same stack_id will be stacked together, allowing for flexible combinations of stacked and grouped bars.View Documentation
✨ Rotate x-axis labels
Use the newlabel_rotate option in x_axis_options to rotate the labels on your x-axis
✨ Models (Beta)
Join sources together and materialize the results to improve performanceModels can be used to execute joins across sources, centralize type casting, and create clean tables to use when building reports. Models are run and materialized every 8 hoursView Documentation✨ Set custom end date for date ranges
Use a dynamic end date for date ranges based on your data or a standard offsetBy default, date ranges end as of today’s date, but this setting allows you to choose a date based on a standard offset (e.g., 1 day ago) or by passing a sql query to calculate the date (e.g.,select max(date) from my_table)🚀 Variable name shortcuts + use in markdown
Variables no longer require specifying a property - simply use the name, like{{ my_var }}You can also use variables directly in markdown, including in headers and lists. You still have access to variable properties and can reference them like this: {{ my_var.selected }}
🐛 Improved date behaviour
Previously, database columns that returned Date type would return DateTime precision, which is confusing in the UI. Now dates will return no additional time componentValues that are Date type will no longer display time component in Tables, Values, and in various input components. Eg a Date previously formatted 2025-09-22T14:05:23Z would now be returned as 2025-09-22.
✨ Page Loading (table)
Tables now load instantly on preview and published pagesThis is the first change in a journey to fully enable Server-Side Rendering in Evidence pages. The table is the first component to support this behavior, which enables the following features:- Tables load instantly on preview and published pages (no more loading spinner when the page loads)
- Chat can now debug SQL query errors in Table components
This change lays the groundwork for other components to opt into this new Server-Side Rendering system.
🚀 Improved Default Funnel Chart Styling
Funnel Charts are more intuitive to read, with a squared off style
✨ Icon Component
Add icons inline into markdown, with configuration for color and size
✨ Row Level Security
Add rules to restrict which rows of are visible to users from sourcesAdd RLS rules to data source columns, and add user variables to determine which values in these columns allow access.View Documentation🚀 Custom Color Palettes
All charts now support custom color palettes via the “chart_options” attributePass an array of hex codes to the color_palette option e.g.color_palette=["#ff0000","#00ff00"]
✨ Link Button
Use the link button to link to another page in Evidence or an external locationView Documentation
✨ Variables
Define variables in frontmatter and use them in markdown, SQL and componentsView Documentation✨ Page linking
Link to other Evidence pages using markdown links✨ Filter interpolation
Reference filter values inside of inlined queries, and view the state of your filters in the new filters pane of the devtools sidebar.View Documentation✨ Reference Components
Usereference_line, reference_area, and reference_point to add reference annotations to your chartsView Documentation