What’s shipping at Pendium.

Autopilot now skips topics it's already covered, onboarding shows what AI actually says about your brand, and the 404s AI agents hit on your site are now visible.

Saved workflows are now a permanent part of Create, and the worksheet is gone for everyone.

Meta pixel event coverage expanded across the marketing site, so ad campaigns now get cleaner conversion signals out of the box.

Vercel sites can now track AI traffic across the whole site, and the Analytics Sync dialog shows live sync status at a glance.

Autopilot now publishes posts on its own judgment, and the brand home page loads section by section instead of all at once.

Site Analytics now shows which AI assistant read your pages most, with annotations you can add yourself to explain what drove a spike.

Admin gets a full audit log, scan health tools, and per-record pages; Pendy can now text first, remember conversations, and handle Google reviews.

Free users can now revisit their onboarding scan, and the Pro guarantee feature lets you lock in a measurement baseline tracked over three daily runs.

Free brands get a dedicated dashboard, Site Analytics splits into focused tabs, and every section now has a built-in guide explaining what it does and why it matters for AI recommendations.

Autopilot now has a full audit trail, free accounts move through a proper onboarding gate, and the /mcp page gets a setup prompt for agent-based installs.

Bulk content creation launches in beta, the scan preview gets a visual overhaul, and signup now starts reading your site before you finish the form.

The "What's next" calendar now shows your score moving in real time, and onboarding got a full redesign that actually explains what Pendium does.

The post editor now works on phones, and onboarding shows a live Autopilot preview instead of drafting a page.

The post editor now has a slash menu for blocks and AI writing tools; plus a fix that keeps Search Console data accurate across your whole site.

Homepage A/B rotation is live across three landing designs, with a fix that stops returning users from losing their website on sign-in.

AI traffic from ChatGPT, Perplexity, Gemini, and others now appears directly in Site Analytics when Google Analytics is connected.

Workflow editing is now live — remove, reorder, or add steps before a run, with two new research options and a live credit estimate.

Brand pages are cleaner and more accurate — placeholder scores are gone, broken rival links are fixed, and GA4 traffic data from AI assistants starts syncing in the background.

Search Console now breaks down impressions by country and device, and Google Analytics can be connected for accounts with early access.

Free plan content limits are now enforced, yearly pricing shows the annual total upfront, and social posting data has been removed.

Create gets per-run header and research controls; Conversations merges into Questions; one plan layout now spans pricing, the picker, and onboarding.

Plan-limit messages now always include the right action button, and brand pages no longer show a guessed website before scanning.

The quality check now lives inside the post editor, brand pages can be claimed and verified, and social media posting has been removed.

New signup flow collects onboarding details upfront, and tracked prompt limits are now per brand instead of shared across your account.

Agency plan card, a tracked "Talk to our team" flow, and the groundwork for Pricing v2 limits across brands, scans, and posts.

Inline AI editing arrives in the post editor, free plan access gets clearer limits, and social carousels gain Instagram/TikTok/X scheduling.

Pendium now runs a daily agent that researches shopping categories in AI search and posts a short video about what it finds.

Ads → Performance now shows real ChatGPT campaign data, and Pendium social posts get brand tagging, smarter audits, and real product photos.

ChatGPT ad campaigns now have real spend controls, live performance data, and clearer budget labeling throughout the Ads page.

Pro owners can now access Ads directly from the sidebar, and chain businesses no longer get a single branch mistaken for their whole company profile.

A rich-text editor for blog posts is now available in early access, alongside an autopilot that publishes daily social content from your research.

Admins can now build fact-checked social carousels from AI research data, sized for Instagram, TikTok, and X.

Owners can now publish their own site changes in one click, and a new Tools section surfaces AI-gap ideas with build specs ready to hand off.

Blog posts now lead with your own brand photos, and the admin panel gets a proper user management page with search, filters, and inline controls.

Draft exports now work with a preview token only, so your site integration never needs to hold a Pendium API key.

Web Presence now auto-discovers business profiles across review and directory sites, and the Visibility report got a significant layout overhaul.

Ask the agent about your own visibility results directly from the report — plus several fixes to how cited pages and questions are displayed.

The Pendium Assistant now lives in the sidebar so you can chat and read your dashboard at the same time.

The Create Content page is redesigned around an agent-first experience, with brand-specific content ideas ready to act on immediately.

Free users can now try Pro for 7 days before paying, and several visibility surfaces now show accurate, relevant data.

Autopilot is now a page you can send people to, and the onboarding now walks you from gap analysis to your first published page in one flow.

New sign-ups get a ready-to-publish draft page on a search gap, and /visibility now shows how your score has moved since day one.

Competitors from your Brand Info list now appear in AI Visibility, and shared reports have been redesigned to match the public site.

Shared AI visibility reports are rebuilt around real answers, and Conversations is back in the sidebar where it belongs.

The admin site build view is redesigned as a compact table, and "View site" now opens the right URL depending on whether your site is live.

Brands hosted on a subpath can now migrate to a new site without losing their custom domain redirect.

Pro brands now control when their site goes live, and FAQ schema is automatic for any post that's actually structured like a Q&A.

You can now move a post off Pendium and onto your own WordPress or Webflow site, with a permanent redirect handling the handoff.

Moving a site to brand-sites no longer risks taking it down if the underlying build isn't ready yet.

The site build notice now appears correctly after a Pro checkout, not just in-app upgrades.

Upgrading to Pro from an agent now kicks off the site build immediately, and old domains no longer get flagged as your own competitor.

Domain Connect now handles stuck unverified subpaths so they don't silently block your brand-site setup.

Relative links on subdomain pages (like docs.example.com) now resolve correctly when a page is recreated.

Chat cards render again after a deployment gap, and relative links borrowed from your main site no longer 404 in AI-optimized content.

Visibility scans now run on different AI platforms depending on your plan, and the docs are clearer about where your site actually lives.

Headless CMS traffic reporting is live, plus Site Analytics now counts homepage visits correctly instead of splitting them across two rows.

Custom domain setup on the Site page now works properly on mobile screens.

Custom subdomain sites can now keep their subpath — every page redirects correctly to its matching page on the subdomain.

Brands running their own website can now send AI crawler visits and pageviews into Pendium, with the same analytics they'd get on a Pendium-hosted site.

Subpath routing for custom-built sites works correctly now, including when setting a subpath as primary.

Brand-sites hosting is now a full production feature: real-time revalidation, per-brand addresses, admin tooling, and tightened identity checks across the platform.

Pendium now manages your site's design automatically, and the old Site V2 section is gone — one less place to look for anything.

Menus now show only what your account can actually use — two options that appeared for accounts without the right access are gone.

Production deploys were silently cancelled on every release for an unknown period; that's now fixed.

The site export API is live, free brands now see generated site previews, and a custom domain routing hole that was polluting Search Console has been closed.

Brand logos that were showing as broken images across generated sites and style guides are now repaired.

Ask the style guide to move your phone number to the header or drop the login button, and every page updates to match.

Agent chat is now free, every page type gets a designed layout, and the site engine behind Pendium's brand sites has been rebuilt end to end.

Local guides now support their own billing key, and the drift gate can record a clean "no change needed" verdict without manufacturing one.

Website Widget embeds now support per-embed layout and explainer settings, so agencies can run different configurations on different pages without overwriting each other.

Local guides now reach search engines on publish, and dense metro areas like Los Angeles return actual provider results.

Citation Sources now connect directly to both your content plan and outreach queue, turning a research row into an actionable workflow in two clicks.

The research guides page no longer crashes on load, and the out-of-credits banner that falsely appeared with it is gone.

Local research guides got a quiet overhaul — better business discovery, fewer dead listings, and more complete AI analysis per provider.

Local service guides now surface real businesses ranked by AI visibility score, and free accounts can do far more without hitting a paywall.

Local businesses are now measured where their customers actually are, and your visibility report opens with the highest-evidence quick wins.

Filter Prompts by AI platform to see per-source scores, and a new dashboard card shows how each platform tracks over time.

Onboarding now tells a cleaner story from day one, and newly created sites no longer show misleading "Published Live" badges on pages the public can't reach yet.

Web Presence now scrapes and audits all eleven connected platforms — G2, Capterra, Trustpilot, GitHub, Substack, Medium, Sitejabber, ConsumerAffairs, and Amazon Storefront — with no placeholder data anywhere.

Partner agencies can now embed a white-labeled AI visibility scanner on their own site and hand clients a ready-to-run coding brief.

Local businesses now get a structured fact sheet from their Yelp and Google listings, and homepage generation pulls from real category and geography research.

Generated homepages are now real, self-contained code; the full platform UI reached its new design system; and a wave of navigation, chat, and onboarding fixes landed.

The chat home now opens homepages and AI Visibility reports beside the conversation, and homepage generation gets a design-review loop before anything goes live.

Visibility scans now measure AI search grounding, and a new Fact Sheets page shows exactly what context your AI agent is working from.

Your brand's asset library is now reachable from the sidebar, and Action Plan sheets load faster and stay current automatically.

The chat agent now reasons from live fact sheets instead of guessing, and the site builder pulls real brand assets into generated designs.

Structured data on agent pages now emits correct absolute URLs, fixing Google Search Console validation errors for custom-hosted brands.

Two new MCP tools let agents read and update blog posts without guessing — plus fixes to credit reporting and tool documentation.

Homepage preview and publish control are now separate, so you can see your page before anyone else does.

The brand logo no longer appears twice on agent sites and hosted blogs when the captured nav bar is shown.

The generative homepage is now available to all brands, with its own sidebar entry and a refreshed onboarding experience.

Pendy's AI agent now runs on its own API key, keeping its usage and rate limits separate from the rest of your account.

The homepage now walks anonymous visitors straight through a live visibility scan in chat, without requiring a login or redirect.

Your agent now catches duplicate content suggestions, SMS cards arrive as images, and social profiles stay clean without manual cleanup.

The chat agent now generates shareable visual cards on demand, and your dashboard recommendations are grounded in what your website and online presence actually cover.

The chat agent now knows what's actually in your knowledge base before it speaks, and can pull in the rest of your site without you leaving the conversation.

The chat agent now has long-term memory — it remembers decisions and preferences across conversations and recalls older ones when relevant.

The Grow page now has an industry picker, so you can see recommendations filtered to your sector without going to a specific industry URL.

The Google Business Profile link on your Web Presence page now points to your actual Google listing instead of a broken Maps search.

The MCP tools now return real, ready-to-share URLs for posts and feeds instead of hand-built ones that broke on custom domains.

Pendy's monitoring card now shows all AI platform logos, company favicons load correctly, and her prompts no longer pull her toward the old scan-and-wait flow.

Pendy's visual cards now render correctly in production, and a cluster of message-logging fixes make its send counts and transcripts accurate.

Pendy now handles images, files, links, and split iMessage threads correctly — no more dropped attachments or duplicate replies.

Pendy cards are now more resilient — numeric IDs, missing fields, and malformed data no longer break renders.

Location fields like "Brooklyn" or "San Francisco Bay Area" now reach the AI correctly instead of being silently dropped or misread.

Location detection no longer misreads country pairs like "US, UK" as a city and state, so local-business classification stays accurate.

Buyer personas no longer get a city attached to national or SaaS brands during onboarding.

Text fields no longer jump the page on iOS Safari, and the Pendy chat channel now routes through Linq.

Agent sites now render headers and footers from structured data instead of cloning your live HTML — no more broken navs, blown-up badges, or stuck menus.

Admins can now pause all scanning activity site-wide, and the hosting wizard finally explains what to do when Cloudflare is blocking your AI crawlers.

The chat agent now shows its reasoning, handles multi-step blog tasks end-to-end, and Business Review replaces the standalone Tasks tab as your main action hub.

The in-app chat agent can now generate a page and publish it in one go, without asking you to wait and come back.

The chat agent now runs visibility scans autonomously and follows up on its own — in-app and over SMS.

Business Review turns your scan data into specific, quoted strategic moves — and a full competitive breakdown is one click away.

Clone theme now reproduces your real nav and footer on any site builder, and Web Presence suggestions come with ready-to-publish copy.

Local businesses now get the right city in their AI profile — no more being reassigned to a nearby metro during onboarding.

Web Presence tab launches with profile audits, AI edit suggestions, and on-demand refresh for Yelp and Google Business Profile listings.

The Pendium Assistant now generates links to your site correctly, even when you're on a subpath or your domain setup has changed.

Agent sites can now mirror your real website's look — nav, footer, colors, and type — with one click.

Subpath-hosted agent sites now serve all pages on your domain, not just the blog feed.

Visibility scores in reports now match the dashboard, and branded prompt ranking works correctly when the AI mentions your brand in prose.

The Sources page Query Type filter now controls the whole page and shares state via URL, so filtered views are bookmarkable and consistent.

Sentiment now appears across Topics, Reputation, and every visibility surface — and AI agents citing your brand now point to your actual website.

Visibility scores now count only recommendation prompts across every view, and platform results that were silently dropped now appear where you'd expect them.

Clicking a tracked prompt now opens a full breakdown of its visibility history, competitor rankings, and raw AI answers.

Recommendation Score and Brand Awareness now report separately, so your dashboard reflects actual discovery performance instead of a blended number.

The Brand Accuracy report is rebuilt with a cleaner layout that makes zero counts visible and reassuring, not absent.

The Sentiment page now shows you where AI platforms are getting your brand wrong — and what to do about it.

The Yelp Study dataset now covers 28 industries and 4,335 markets, with live AI surface data from ChatGPT, Perplexity, and Google AI Mode.

Agent sites now show your brand in the nav, and the in-app chat agent can now pull deep analysis from any knowledge source.

The Sources page and Brand Index load noticeably faster — here's what changed under the hood to make that happen.

The Sentiment page loads significantly faster, and the Yelp advertiser study has been rebuilt with cleaner data and honest findings.

The agent chat can now answer questions about your account, and brand extraction is more accurate when your query spans multiple industries.

The in-app agent now requires your confirmation before it creates, edits, deletes, or archives anything — nothing runs without an explicit click.

Brand mention detection now correctly recognizes names when AI responses reformat or add words between parts of your brand name.

The Credits dialog on shared agents now shows the real remaining balance instead of 0/0.

The Outreach page is back up, and teammates on shared agents now see clear messaging about where credits and plan limits actually live.

The Ads page now tracks real campaigns, recommendations are replaced by a live Tasks board, and team members can fully manage shared agents.

The in-app agent can now manage blog posts end-to-end, and the Yelp Ads panel shows your real category ranking and top competitors.

The in-app agent can now generate workflow post ideas from scratch, and adding a new agent no longer risks a blank screen on sign-in.

Ads gets a cross-platform dashboard and campaign builder preview; MCP agents can now take an agent-site page all the way to live.

Local businesses with a Yelp presence can now access the Ads section, and re-indexing a knowledge source no longer risks wiping it mid-refresh.

The Ads feature is live in beta, and agents can now write — not just read — your brand's voice, goals, knowledge base, and settings over MCP.

MCP tools now let external agents manage workflow structure directly, and blog posts no longer publish with broken plain-text URLs in prose.

AI responses that cite a brand as a buying source without naming it in the text now count as a mention in your visibility data.

Credits now show top-up and plan balances separately, and legacy subscribers stopped missing their monthly credit reset.

Cold-email scans now run fully under your agency's brand, and several billing issues that silently shortchanged users on credits are fixed.

Scan preview requests with isPublic set to true now complete successfully instead of returning a server error.

Shared links to agent pages, feed posts, and blog posts now show a branded preview card instead of a blank embed.

Service-area pages are now discoverable from your Services index, and sharing any agent page link now renders a branded card instead of a bare URL.

Partner scan integrations that started 400ing after last week's deploy are unblocked, and Yelp/Google URL onboarding no longer leaves businesses nameless or stalls for a minute.

Yelp and Google Business onboarding now completes in seconds instead of minutes, with fuller brand profiles on the first pass.

Onboarding via Yelp or Google URL now correctly shows your business name and details before asking you to confirm them.

Adding a business by Yelp or Google URL now populates the company overview correctly instead of returning a blank profile.

Service Areas pages now surface correctly to AI agents, and Yelp review capture works again for businesses onboarded by URL.

Local businesses now get structured data, hyperlocal content generation, and per-neighborhood landing pages — all built from the service areas you've already set up.

Customer reviews are now a first-class knowledge source, and the agent site link surfaces everywhere partners and API callers need it.

Google Business Profile now works everywhere Yelp does — preview scans, full accounts, and cross-platform discovery that finds both listings automatically.

MCP agents can now iterate on existing posts end-to-end, and multi-team partners can control which agency identity appears on prospect reports.

The Yelp page now shows live preview scans alongside full brand reports, so visitors can see real examples at both depth levels.

Yelp-seeded brands now show their real logos and share links, after a domain collision bug was sending visitors to the wrong business reports.

Partner custom domains are now shared across the whole team, so any member can deliver white-label reports under a single agency domain.

White-label reports no longer expose Pendium branding, and three silent cross-tenant write vulnerabilities in brand settings are now closed.

Agent pages are now connection-rich by default, the scan paywall has a no-card entry path, and local-business brands get a directory-style agent site.

Three new developer guides land alongside a fix that stopped a scraper from taking down the database by making public blog pages cacheable.

Agent site roots now serve your homepage directly, X integration is hidden until the API situation resolves, and new customers get a founder welcome email on signup.

You can now replace blog header images with your own uploads or stock search, and enterprise accounts can scan across more AI platforms and model variants.

The Visibility explorer has moved to Personas and Topics, giving each page a focused exploration experience.

The MCP server and REST API now accept your personal API key as a bearer token, so you can connect from tools that don't support OAuth.

Databricks MCP connections now authenticate and persist credentials correctly through Unity Catalog's per-user credential store.

The in-app AI assistant no longer crashes mid-conversation when you have a lot of scan history.

Trailing-slash URLs across brand pages, admin, and settings were stuck in a redirect loop — that's fixed.

Post actions (retry, regenerate, repurpose) now enforce tenant access correctly, closing a gap where those operations didn't verify ownership.

The Posts page gets bulk scheduling, the Feed view shows what AI agents actually see, and tracked prompts are easier to organize.

Prompts added from in-app surfaces are now rewritten into natural buyer-intent queries, plus partner branding settings got a full consolidation.

AI agents can now discover Pendium's API and MCP server automatically, and analytics queries run more reliably under PostHog load.

Blog analytics no longer times out when refreshing — pageview and referral data now loads correctly across all tracked sites.

Agent sites now expose machine-readable discovery endpoints for AI crawlers, and a settings reorganization makes site configuration easier to find.

The account journey trend now reflects real movement, not sampling noise — plus a security fix for shared report actions.

Visibility scores are easier to understand, the y-axis on trends charts now reflects your actual data, and agent site hosting controls have a permanent home.

Visibility trends over time, a smarter scoring formula, and deeper scans that stay within budget.

The account-review deck and Site Analytics now surface conversational AI traffic — the visitors who actually generate recommendations — instead of raw crawler counts.

The account-review deck now leads with prompt-level insight instead of an overall score that drops by design when you add harder prompts.

Tracked Prompts can now be seeded from real search data, and Site Analytics traffic numbers are more complete and accurate.

Site Analytics now breaks down traffic by humans vs. conversational AI agents at the page level, so you can see exactly which content AI is fetching for real users.

Agency partners can now serve client reports from their own domain, and AI bot traffic in Site Analytics is split into three meaningful categories.

The Account Review deck turns your scan analysis into a slide-by-slide briefing, and two new content-engineering pages open a second conversion funnel.

Partner profile URLs now use a direct indexed lookup, making branded links faster and setting up custom domain support coming soon.

Per-prompt visibility scores on the Topics page, a public agent-site preview route, and a fix for blog posts generating without titles.

Autopilot no longer batches up missed scans as a debt to repay — missed cadence is forgiven, not owed.

Workflow worksheet now handles deleted posts gracefully, and subpath-hosted agent sites no longer serve broken JS or images.

Every scan now tells you what moved since the last one and why — scores, citations, competitors — with suggested next steps.

The workflow outline editor now surfaces all 13 structured fields, and a new admin tool lets you watch AI-powered site previews assemble in real time.

Subpath proxy setup was silently broken for all customers; both that and malformed credit labels are now fixed.

Feed Health moved to admin-only, and the KEY FINDINGS card on brand reports now shows actual findings again.

Google Search Console verification now takes one paste, plus a new AEO scorecard shows exactly how well your agent feed is set up.

Slug editing got a full overhaul, partner scan URLs now redirect cleanly, and several silent data-loss bugs in the workflow editor are fixed.

A security gap in post asset updates is now closed — cross-tenant requests and missing post IDs are handled correctly and indistinguishably.

The subpath hosting wizard now generates real reverse-proxy snippets, and partners get branded scan preview URLs they can hand directly to clients.

Agent view of private blog posts now works correctly for brand owners, plus a security fix for the owner bypass on public routes.

Manual competitors now show mention counts from live scan data, and white-label agency guidance starts with real default content instead of a blank field.

Platform admins can now manage group-level white label branding directly, without needing DB access or group admin rights.

Shared visibility reports can now carry your logo, name, and strategic voice — no plan upgrade required.

The chat agent can now manage the full content lifecycle, and the Visibility Explorer closes its biggest gap with the Latest Report.

The chat agent can now edit your brand kit directly, and next-action recommendations stay consistent across the dashboard, fallback logic, and emails.

The in-app chat agent can now trigger a visibility scan directly from the chat sheet, without switching screens.

The in-app chat agent can now manage your brand's agent pages directly — create, edit, publish, and delete without leaving the chat.

Docs links that broke in a recent rename are fixed, and the in-docs search bar works again.

URL inputs across every public marketing page now start the same lightweight visibility preview that the homepage uses.

The research pipeline now runs end-to-end without manual intervention, publishing roughly three new articles per day automatically.

The in-app assistant can now edit your blog posts and manage your knowledge base directly from chat.

Scan preview links are shorter and any old URLs you've shared redirect automatically to the right place.

The in-app assistant can now read your blog, manage workflow rows, and link directly to the right page — plus a new Competitors view and a renamed Recommendations section.

Bookmarks replace the old Analyzed URLs tab, scan-complete emails now open with an AI-generated brand summary, and insights buckets derive from card polarity instead of a separate LLM call.

Recommendations now come from the same LLM call that produces insights, so the two stay in sync and trace back to each other.

The dashboard now surfaces a single AI-generated next-step card, and demo leads get a follow-up email when their report is ready.

You can now hide your agent site's feed without taking down the rest of your pages, and site analytics has a clearer home.

Internal test scans no longer trigger Slack alerts in the production channel.

The Setup Guide moves into the sidebar, and URL-seeded workflows now check your credit balance before running.

Gemini scan failures caused by invalid model IDs are resolved — research mines and batch scans now complete reliably.

Cold-email outreach reads more like a real person wrote it — better discovery framing, cleaner subject lines, and guardrails against hollow observations.

Research Hub articles now show real brand scores, correct domain links, and properly formatted content — plus a full visual and layout overhaul.

The cold-email generator now leads with opportunity instead of calling out what's broken — better for opens, better for client relationships.

Cold emails now render as short, scannable paragraphs and brand grounding fires automatically on every generation.

Production deployments were silently failing mid-build; that's fixed.

The sidebar is restructured around a Monitor/Analyze/Grow loop, agent pages get a unified layout, and a full cold-email pitch toolset ships for partner teams.

The free plan no longer shows the 150k credits item on the pricing page.

Visibility Scan Previews now include a competitor rankings slide that shows where you stand head-to-head against others in your category.

Conversion events now fire server-side so a closed tab can't drop a signup or upgrade from your Meta attribution.

Agent feeds are now fully crawlable by AI bots, and logged-in routing gets smarter about where to send users based on what they actually have access to.

Conversion tracking now fires one clean event per action — signups, purchases, upgrades, and top-ups each have a distinct name across Meta, Slack, Loops, and PostHog.

Pricing page and in-app subscription copy now more clearly describes what each tier actually does.

The Visibility Scan Preview is now accessible via API and MCP, so outreach tools can trigger and read full brand analyses programmatically.

The Shopify industry page now leads with clearer copy and defaults to Shopify when you arrive.

Shopify storefronts now scan correctly and display cleanly across the product — two bugs that quietly broke a lot.

Pendium now understands Shopify stores — point it at any storefront and it pulls your catalog, best-sellers, and collections automatically.

The bot-activity badge on brand pages now shows real vendor logos and stays out of your way on every screen size.

The "AI agents read this brand" badge on your brand page now shows current activity instead of a number that could be up to 24 hours old.

Brand Index pages now show a live feed of AI agents visiting them, and crawler details across the product got meaningfully clearer.

The About page is rebuilt around the founder story, with all five investors now listed and the product diagram removed.

The Feed Analytics empty state no longer tells you to connect Google Search Console when you already have.

Reddit outreach drafts now work reliably in production, and the import-to-account flow no longer fails after fresh signups.

Public Visibility Scan Preview pages now show the site header and footer, and handle server errors gracefully instead of showing a blank Vercel page.

The public Visibility Scan Preview page is now resilient to incomplete data and broken favicon URLs.

Vercel log drain noise is gone — production logs now only write when there's something worth seeing.

Visibility Scan Previews are now public, indexable pages — turning every scan into a shareable snapshot of how AI platforms see a business.

The "Visit Site" link on agent pages now opens the right URL instead of a broken Pendium path.

Marketing pages load faster on first visit, and visibility scans now run on updated model defaults.

AI Agent Activity now updates within minutes, not hours — traffic you send today shows up on the dashboard today.

Agent feed visits are now ingesting correctly after a silent data loss bug, and public pages load faster from edge cache.

You can now see exactly which AI crawlers are reading your feed — GPTBot, ClaudeBot, PerplexityBot, and others — directly in Feed Analytics.

The weekly AI bot list sync no longer crashes when the upstream feed contains duplicate entries with different casing.

Feed Analytics is gaining AI agent tracking — this release lays the data foundation for seeing which AI bots are visiting your feed and when.

Demo landing pages are now indexable by search engines and load faster on mobile.

The demo flow is now two steps instead of three, dropping users directly onto their scan report with the booking modal open.

Content Opportunities surfaces your real GSC search demand inside Agent Feed Analytics, turning ranking gaps into AI-optimized content ideas.

Edge CDN caching now works for public pages, worksheet research gets better coverage, and blog posts no longer save half-written.

Feed Analytics now separates your search and engagement data cleanly, and AI suggestion tools stop recycling the same personas and topics.

URL analysis pages now end with a direct path to action, and workflow ideas can be regenerated when the first angle misses the mark.

Legacy blog URLs with double-hyphens now redirect correctly instead of returning 404s.

Blog post URLs are now clean and stable, and URL-seeded workflows no longer run automatically on creation.

The onboarding scan page now shows your actual data as it loads — business info, personas, and live AI scores — instead of placeholder content.

URL-seeded workflows now show the source page and grounding status inline; onboarding scan now lands correctly after login.

Analyzed URLs can now seed content workflows directly, and the homepage scan flow has been rebuilt to capture more visitors before they bounce.

Worksheet notes, URL-seeded workflows, and a re-analyze action give you more control over how content gets generated.

Autopilot posts no longer use verbatim seed titles, and you can now control post length directly in Agent Feeds workflows.

Onboarding scan funnel now fires Slack alerts and Meta Pixel events at each stage, so you can see exactly where prospects convert.

The onboarding scan page now matches the marketing site design, so new users land in a consistent experience from the first click.

A new pre-signup visibility scan lets prospects see how AI platforms perceive their business before they create an account.

Onboarding now stops early on invalid URLs instead of wasting a full scan on parked domains, login walls, and other junk.

Feed category filtering now works correctly end-to-end — from URL to server to newly generated posts.

AEO blog content gets a major quality overhaul, and the onboarding workflow now teaches the app by doing.

Outreach target search now shows a live "Classifying…" state for the full job duration instead of snapping back to idle after a second.

Outbound links on your public agent pages and feeds now carry a UTM parameter so you can see Pendium-referred traffic in your analytics.

Workflow idea generation now uses your full brand profile and knows what you've already published, so it stops suggesting content you've already made.

The Outreach feed turns your visibility scan results into ready-to-send emails and pitches, drafted in your brand voice.

The Citation Sources page has been rebuilt around how AI platforms actually cite your brand — with smarter classification, partner tracking, and manual overrides.

The Visibility report opens on Full Report by default, and the Visibility Landscape now shows the actual query prompts and competitor context in one place.

Google Docs is now a knowledge source, and a round of fixes cleans up autopilot, knowledge processing, and the Visibility Landscape table.

Workflow credit estimates are now visible before you run anything, so you know what a full sheet costs before clicking Run All.

Credit balance now stays live during background jobs, and several setup gaps that could cost you indexing progress are closed.

Key Pages gives every outline a real internal-link plan, and content generation now has strict anti-pattern rules that cut AI-sounding output across the board.

Your agent feed and profile pages now go live automatically on first setup, so new accounts start fully visible to AI platforms without manual publishing steps.

Blog articles now generate in full, and onboarding scans no longer publish brands to the public index without manual approval.

Branded 404 pages now serve the right content to AI crawlers and humans alike, instead of broken or misleading responses.

Company profile is now a single shared record across Brand Kit and Pages, and onboarding fills in richer AI-context fields automatically.

You can now run agents in languages other than English, with generated queries, personas, and posts all following suit.

Priority Level grouping is now the default view in Visibility Explorer, and content generation now matches your website's language.

The Sources tab now defaults to 90 days, the Ideas column is directly editable, and Feed Settings picks up an Offerings Label selector.

Autopilot now runs a full content workflow per post, and you can re-trigger brand discovery from the Overview page without clearing your existing data.

Onboarding now starts with tighter, more targeted queries — and admins can trigger autopilot runs on demand without waiting for the scheduler.

Public release pages at /releases give every version a permanent, readable URL — and "Run New Scan" now takes you somewhere useful.

Per-agent autopilot is live — schedule visibility scans and blog posts on a cadence, with live credit estimates and a rebuilt agent dashboard.

Post exports now bundle the asset with the markdown, the AI/humans toggle looks the same everywhere, and /{username} always lands somewhere real.