All releases
v1.9.0

Agent Journeys, Instrument Redesign & Source Influence

Agent Journeys sends an AI agent through a real task on your site and traces what it did, alongside a new flat visual design across the app and marketing site and Source Influence to see which sources AI models cite for your brand.

Features

  • Agent Journeys — Send a browser agent through a real customer task on your site and get back a step-by-step trace of what it did, where it got stuck, and what it concluded. Choose the model that drives the agent (ChatGPT, Claude, Gemini, Grok and others), build a run from a short form, and watch it execute in a live console.
  • Source Influence — See which sources AI models actually retrieved and cited when they talk about your brand, now with a per-provider drill-down showing how each model's sources differ.
  • Redesigned App — The entire authenticated app has been rebuilt on a new, sharper flat visual design that loads faster and drops an old stylesheet that had been fighting the theme.
  • Background Agent Runs — The AI assistant keeps working when you navigate away, with a persistent header showing your active run and recent chats.
  • Changelog — A public changelog, so every release is visible without asking.
  • Apps Settings — The Apps page has been reworked with a Browse tab and clearer app lists.

Improvements

  • Sidebar Icons — Every section of the sidebar now has its own unique icon, replacing the old shared icon set.
  • Dark Mode Refinement — Adjusted dark mode colors for better contrast and a more consistent look across the app.
  • Redesigned 404 & Loading Screens — Error and loading pages now match the app's new visual design.
  • Marketing Site Refresh — The public website has been rebuilt to match the app's new look.
  • Honest Discovered Domains — Discovered Domains now distinguishes a domain that has never been checked from one where your brand genuinely wasn't found, instead of showing both the same way.
  • Clearer Source Wording — The brand co-mention figure now says what it actually measures, and mention counts make clear when a source was retrieved but not cited.

Fixes

  • Add-Market Button — Fixed the Add Market control disappearing from the Share of Voice overview page when an account had an open market slot.
  • Account Switching — Fixed the AI assistant carrying over a previous account's conversation and chat history when switching accounts in the same session.
  • Dropdown Menus — Fixed dropdown lists extending off-screen or collapsing to a sliver on short or zoomed-in windows.
  • Knowledge Base Counts — Fixed document counts and status showing stale values in the knowledge base list.
  • Missing Journey Findings — Fixed a journey run discarding every finding it produced, caused by an internal version label that was too long for the field storing it.
  • Inflated Discovery Counts — Fixed our own index and unverifiable retrievals being counted as discovered domains.
  • Share of Voice Timeouts — The per-provider time limit now applies to each provider separately rather than across the whole batch, so one slow model no longer cuts the others short.
  • AEO Audit Parsing — The readiness audit now reads structured data that contains raw control characters, uses @graph nodes, or wraps values in arrays, instead of skipping the page's markup entirely.
  • Screenshots — Fixed page screenshots failing with a server error.