• Watch the AI's reasoning as it works, and see each task it changes
• AI agent edits can now be redone after an undo

Notifications

• Outstanding study work shows as badges in context

• Insights now splits into Summary and Related themes tabs.

Participant Credits

• Every customer can now see the credits log.

Fixes

• Failed Figma imports now say what actually went wrong.

• Themes replace the findings panel in the video viewer

• The AI Moderator is now generally available on live-site tests — no longer behind a staff flag.
• Live captions stay in order per step, even when a participant talks over the moderator.
• The moderator's replies sound more natural — more varied acknowledgements and fewer robotic "You're saying…" restatements.

AI Moderated interviews

• Interviews now connect reliably, even on strict office networks.
• The AI politely waits for participants to finish answering its final question before wrapping up.
• Spoken answers to open-text questions now show up as text in your results.
• Voice quality got an upgrade across the board.

Insights & reports

• New live-site details modal — slice moderator Q&A by question or by participant.
• Drag Themes straight into a report from the sidebar.
• Report notes and themes now stay in sync as live snapshots that refresh automatically, and show "Source deleted" when the underlying task is gone.

Moderated sessions & playback

• Moderated sessions listing, with livestream player fixes.
• Native keyboard controls for seek and play/pause in the player.
• "Open in Hubble" button on shared recording viewers.

Fixes

• In-product surveys & guides: a trailing/*URL rule now also matches the bare path.
• Recruitment statuses simplified — "completed" retired from the status vocabulary.

• Participant statuses now match everywhere you look — study pages, the Share tab, and Overview all tell the same story.
• Overview numbers load noticeably faster, and show a dash while they're on their way, so you're never looking at a misleading zero.

AI Moderated interviews

• Interviews now connect reliably, even on strict office networks.
• The AI politely waits for participants to finish answering its final question before wrapping up.
• Spoken answers to open-text questions now show up as text in your results.
• Voice quality got an upgrade across the board.

Fixes and Polish

• Unlock a published study and start editing instantly — with safeguards that keep your live data safe.
• Importing a study plan is smoother, with clearer guidance on supported formats.
• Participants taking multiple studies on the same device no longer get their answers mixed up.
• Dozens of small touches: tidier lists, cleaner modals, smarter defaults.

• Create studies and study plans straight from Claude/ChatGPT
• Claude can now read your existing study plan and asks better questions about your research goals before building, so generated studies match your intent more closely.

Recruitment & participants

• Participant statuses are clearer and more accurate: "Completed" is now "Paid", cancellations made on the Respondent panel side sync automatically, and no-shows are surfaced on canceled sessions.
• The recruitment listing loads faster and metrics are more reliable.

Moderated sessions

• Co-hosting and host take-over: invite a teammate to co-manage a session or take over as host mid-flow.
• Calendar privacy: personal calendar event names are hidden from teammates — others just see "Occupied".

Summary & Findings

• Fixed finding titles and recording summaries to open with the keyword

Fixes & polish

• Published unmoderated studies now lock their editable fields to protect live data.
• Live-stream playback errors are shown instead of a blank screen.
• Smoother sign-in: no more stray error toasts or unnecessary redirects to login.

• Made System/page events easier to distinguish from user actions in the click sequence.
• Made the “Reached” milestone stand out more so success points are easier to spot.
• Fixed timeline labels so they use the available width before truncating.
• Added the foundational model/config support needed to build AI Moderator Mode for live-site studies.

Thematic Analysis

• Added background shade to the newly generated themes.

Global Renaming

• Added global renaming to all study types

Summary & Findings

• Fixed finding titles and recording summaries to open with the keyword

Minor Fixes

• Added loading UI to the moderated study session tab.

• Merge themes with an AI-prepopulated title & description.
• New "Reframe with AI" action on theme detail.
• Noteworthy themes are now surfaced automatically.

Live-site Testing

• Redesigned tester detail as a clean vertical timeline.
• AI summaries per participant and per task in the results modal.
• Drag per-participant AI analysis into your report sidebar.

Notes & labels

• Faster, lightweight label dropdown on notes.
• AI suggests note labels — nested under the right parent automatically.

Scheduling

• 15-minute granularity on moderated availability; new 45-minute session option.
• Rename a participant once and it updates everywhere.

Sign-in and Workspace

• Self-serve SSO for enterprise workspaces.
• More reliable sign-in (new secure session model) and smoother invite/onboarding flows.
• Reorganized Settings into clearer sections.

• Improved participant identity matching across studies.
• More consistent participant tracking throughout the recruitment lifecycle.
• Better display of participant information in results and recordings.

Privacy & Security

• Automatic redaction of personally identifiable information (PII) in uploaded screenshots.

Study Management & Reporting

• Improved study summaries, interview views, and insights displays.
• Enhanced interview metrics and recording management.
• Streamlined study setup with improved templates and interview configurations.

Miscellaneous

• Theme and discussion-guide UI polish

• Redesigned bulk actions on the Notes tab

AI Moderated Interviews

• Preview AI moderator voices before you choose one

Participant Recruitment

• Returning participants recruited via User Interviews are now recognized automatically

Study Setup

• Set a custom label for the "Other" multiple-choice option
• Add researcher contact details to the moderated booking screen
• Use custom tester IDs when sharing a study link
• External link task is being deprecated

Miscellaneous

• Resume a paused study without re-running publish validation
• Theme and discussion-guide UI polish

• Enhancements to AI Note Generation Options
• Improved semantics and note generation styling

Participant Recruitment UX

• Moderated Respondent length locks to the selected study's sessionLength
• User Interviews supports shorter task length
• Recruitment from a moderated study auto-fills length

Unmoderated Study Builder UX

• Aligned header buttons, unarchive support, correct "Resume" label on paused studies
• Resume is its own action — separated from publish in study and survey builders
• Settings button moved onto the in-product survey insights header
• Renamed test-type and question-subtype labels for clarity
• Pulse indicator (recording, push-to-talk) componentized into a shared UIPingDot

MCP Server

• External agents using MCP can now create and update recording notes — same access rules as REST writes. Read tools also surface the persisted transcript quote for each note.

• Dynamic length picker on UserInterviews + Respondent so session duration matches your panel pricing tier
• Past-participant badge now shows across all statuses, with current-recruitment filtering
• Hourly screener-applied notifications + cached Respondent project lookups for tighter copy
• Notifications on participant-initiated reschedules

Video viewer + Notes

• Arrow-key seek and hardened play/pause keyboard shortcut
• Bigger timestamp labels for legibility
• Click a timestamp range to acknowledge a pending note
• Auto-focus the title input on newly-created labels
• Stable note ordering across reorderings (array-index keys replaced with stable IDs)
• Add citations that open directly in edit mode

Pre-made Intros

• upload your own introduction video instead of recording in the builder. Dropzone validates mp4/webm.

Thematic Analysis

• Friendlier empty-state UX for theme generation
• Skips the run when fewer than 2 recordings have notes (no more "1 theme from 1 recording" nonsense)
• AI stops suggesting labels when one already matches
• AI honors custom speaker names + guards against non-JSON model output

Other Miscellaneous

• Shared recording notes now hide source + participant identity
• Notification row polish (compact icon variant, empty-state copy, full description on hover)
• Allow email notifications
• Workspace settings persist correctly when moving between tabs
• Workspace default styling for studies

• Improve MCP tools with URLs, labels, multi-team support
• Auto-generate notes from custom, moderated, and prototype recordings
• Set time expectations before bulk note generatio
• AI summaries for recording-based and prototype tasks

Video viewer

• Add speaker renaming
• Improved note card layout and select-all behavior
• Smaller player controls in tight spaces
• Visual citation range markers on the video timeline
• Drag markers to adjust citation boundaries
• Add citations that open directly in edit mode

Reports

• New sidebar cards: prototype paths, desktop tests, mobile tests, video questions, and AI interviews
• AI analysis on recording and prototype task cards
• Smaller player controls in tight spaces
• Visual citation range markers on the video timeline
• Drag markers to adjust citation boundaries
• Add citations that open directly in edit mode

Settings

• MCP integration card
• AI and recording preferences tabs

• Label template support for faster tagging and organization
• Enhanced Webview study settings, including deeplinks, app download split, and internal staff features
• Improved video player controls for better usability in smaller viewports
• Added border animation to AI note generation button

Improvements

• Updated report cards for Mobile Test, AI Interview, and Video Question
• Discussion guide presets now default to the Custom option

• Moderated study live streaming
• Bulk AI Note generation
• Task groups in summary page

AI Moderated Interviews

• Upgrade model version and improved usability
• Discussion guide presets now default to the Custom option

Report Editor

• Title and nav moved into the editor card; public report viewer border removed and scrollbar repositioned

Bug Fixes

• Several minor bug fixes
• Discussion guide stays clickable when study is locked
• Fixed delete table crash and chart data mutation on navigation

• Feat: add public video viewer page for share links
• Feat: add public note viewer page for share links
• Feat: add public report viewer page for share links

Figma Integration

• feat: add reconnect Figma account flow and fix validate error handling
• fix: prevent auth screen flash during Figma reconnect

Report Editor

• feat: replace static chart images with live ChartNode in report editor

General Improvements

• feat: add compact size variant to UIModal for simple confirmation
• feat: Sync UI/Respondent update with Recruitment form update

• Branch wrap-up prompt based on the participant's last task
• Pass previous task responses to AI voice interviewer for context-aware follow-ups
• Skip ice-breaking for moderated/custom recordings
• Home listing — Changed filter storage from session to local storage so selections persist across sessions

AI Summary & Findings

• Make AI summary output more concise
• Make finding titles concise and actionable

Unmoderated Studies

• Add alert when leaving page during an unmoderated session
• Sidebar UX improvement for unmoderated summary
• Fix button skeleton size on unmoderated insights
• Fix hallucination on Asian transcription and user answer fabricating behavior
• Show 404 page when unmoderated study is not found

Prototype Testing

• Restructure i18n keys for readability
• Minor style improvements

Bug Fixes

• Skip responded check in preview mode to prevent Thank You screen persistence
• Pass study prop to UnmoderatedTask in AI review modal to display task labels
• Add feature label to RunDetail
• Fix study duplication with task group set

• Study access control - Added access levels and collaborator management for studies, with access control enforced on both list and individual study endpoints. Viewer role locks UI controls across builders and reports
• Cookie guide - Added allow-cookie guide for prototype testing
• Video viewer — Fixed active transcript block not tracking video playback, processing state not showing for AI summary and transcript, and note modal state not resetting when reopening
• Report viewer — Fixed back-navigation runtime error and content not persisting in UI
• In-product builder — Fixed showing Publish instead of Pause for live studies
• Home listing — Changed filter storage from session to local storage so selections persist across sessions

• Prototype result card redesign
• Workspace role management
• Unmoderated task group improvements
• Unmoderated tester view polish
• Other miscellaneous changes

• AI-generated notes from recordings
• Prompt support for AI note generation
  
• Voice-only response option in video questions
• Password change support in personal settings
• SDK installation script with team ID in publish flow
• Tester ID tooltip for staff users in responses view
• Video analysis improvements
• Enforced required media questions in recruitment setup
• Disabled actions on archived or completed recruitment projects
• Improved CSV export for study and survey responses
• Improved consistency between studies and recordings
• UI fixes across ranking, card sort, tree tasks, navigation, buttons, and typography

• [Video analysis] Minor improvements
• [Moderated] Minor bug fixes
  
• [Recruitment] User Interviews deadline logic fix
• [Recruitment] Add guard for duplicate tester
• [Recruitment] Allow direct invite from unqualified
• [Recruitment] Improve User Interviews tester qualification logic
• [AI Analysis] Add finding / citation CUD
• Consolidate rename functionality

• [Unmoderated] Prototype mobile view improvement
• [Video Analysis] Video analysis improvement
  
• [Moderated] Microsoft Teams Integration
• [Moderated] Custom links for video conferencing

• [Recruitment] Minor style improvements
• [Recruitment] Allow target participant number changes
  
• [Video Analysis] Add chapters
• [Video Analysis] Add auto-scroll and other improvements

• [Moderated] Add rescheduling / cancellation
• [Moderated] Add confirmation modal for booking details
  
• [Moderated] Add chapters
• [Video Analysis] Add auto-scroll and other improvements

• Improved logic jumps to handle based on task ID versus index
• Logic jumps will be preserved even after changing task orders
  
• Improved UX with ability to specify study length
• Added preset options for AI study generation

• Added full study recording support, including all tasks
• Full video recordings will be downloadable
  

• Added a randomizer in unmoderated studies to allow shuffling of tasks or randomly selecting one from a list of selected tasks to eliminate bias

• Added MP4 Video support for Safari and WEBM for Firefox
• Separated audio and screenshare permissions
  

• AI unmoderated study generation

• Updated Figma's integration to Embed Kit 2.0

• Added image support for card sorting & ranking tasks

• Option to target devices (mobile vs desktop) for unmoderated studies
• Fixed rendering order for open text cards in summaries
  
• Added device tags for home cards