Flight Director is now available for Flight Engineers!

What is Flight Director?

Flight Director is our new agentic AI search engineer inside Market Brew. It replaces Calvin and is one of the most advanced AI SEO agents in the world.

This is a major step beyond a normal chat/RAG assistant. Instead of only answering questions from docs, Flight Director can inspect a customer account, review Market Brew data, understand the configured website strategy, watch model signals, and recommend or even configure the next concrete actions an SEO engineer would normally have to figure out manually.

Things to try

 • “What should I do next?”
 • “Should I create more Ranking Sensors?”
 • “What Flight Plans should I create?”
 • “What page changes would you recommend?”
 • “How do I improve AI Visibility?”
 • “What should I do for LLM Optimization / GEO?”
 • “Which Market Brew signals look weakest right now?”

When Flight Director needs to inspect account data or configure data for your account, it will kick off a background runner. You can watch progress in Flight Director Logs, and the final recommendation / execution results will appear back in the chat.

Why this matters

Flight Director targets one of our biggest operational blind spots: turning Market Brew’s data into clear, expert SEO action.

Today, customers often need a trained Market Brew SEO to interpret Ranking Sensors, Flight Plans, Visibility Launchpad, Listen, Keyword Fueler, Content Booster, Brand Briefing, and task outputs. Flight Director is the first version of an active helper that can do that work directly inside the product.

Flight Director can become the customer’s “AI search engineer in their account.”

It helps answer:

 • What should I work on first?
 • Which pages need attention?
 • Which signals are weak?
 • Which Ranking Sensors or Flight Plans should I create?
 • What exact page/content changes should I make?
 • How do I turn Market Brew data into AI search visibility gains?

This is a strong story for customer success and onboarding:

 • Customers get expert guidance without waiting for a Market Brew specialist.
 • Flight engineers can scale their work across more accounts.
 • Market Brew becomes not just the system of record for AI/search signals, but the active operator that tells teams what to do next.
 • This moves us toward selling Market Brew as an AI SEO operating system, not just analytics.

Before testing: configure the website strategy

Flight Director uses the new Flight Director Website Strategy section in Crew Settings.

For each licensed website you want Flight Director to analyze, fill out the Flight Director strategy fields. This is how Flight Director knows which sites are in scope and what business/search strategy it should optimize toward.

The key fields are:

 • Website strategy / goal
 • Hero keywords
 • Hero landing pages

Important behavior:

 • If a website does not have Flight Director strategy filled out, Flight Director should not act on that site.
 • If multiple sites exist, Flight Director should only focus on the sites with strategy configured.
 • If the user asks for account-data recommendations and no strategy is configured, Flight Director should ask them to complete Flight Director Website Strategy in the Crew Settings first.
 • Generic “how do I use Market Brew?” questions should still be answered without requiring strategy setup.

This is still early, but it’s the shape of something big: Market Brew data interpreted by an agent that knows our workflows and can tell customers exactly what to do next.

Smarter URL Summarization in Content Booster

• The Ignite/Re-Ignite button now enables immediately after entering a valid keyword or focus prompt.
• Validation and mode messaging now refresh live while typing.
• Alternate input modes are disabled sooner to prevent keyword and focus prompt conflicts.
• URL summary and short-prompt helper actions now keep the Ignite state synchronized automatically.
• The global loading animation has been reduced in size for a less intrusive experience.

Content Booster now summarizes URLs more reliably inside the Focus Prompt workflow, including pages that redirect, initially return empty content, or require standard browser-style request headers. Summaries are also cleaner and more focused on the actual article or primary page content instead of navigation, layout elements, or surrounding page chrome.

Additional improvements include:

Additional improvements include:

Content Booster now reacts immediately while typing target keywords or focus prompts. Validation states, Ignite actions, helper options, and prompt/keyword mode changes now update in real time without requiring users to click outside the field first.

• Improved Focus Prompt URL summarization reliability for redirected or partially loaded pages.
• Better compatibility with websites that require browser-like request headers before serving content.
• Cleaner extracted text that prioritizes readable article and main-page content before summarization.
• Redirected Focus Prompt URLs no longer preserve keyword or prompt text in the browser URL when no content item is selected.

Faster, More Responsive Content Booster Input Experience

🚀 Visibility Launchpad, Prompt Noise Management & Performance Improvements

Prompt Noise Management Now Available

Market Brew now gives teams direct control over filtering unwanted prompt activity across both Visibility Launchpad and Ask Widget Conversation Activity.

✨ New Prompt Noise System

A new Prompt Noise management page allows you to:

• Create noise entries per website
• Edit existing entries
• Bulk-delete noise entries

Prompts are normalized before comparison (whitespace and punctuation are ignored), ensuring variations of the same unwanted prompt are consistently filtered.

Any prompt matching a noise entry is permanently excluded from future prompt opportunities.

Visibility Launchpad Improvements

🧹 Manage Noise Directly from the Prompts Tab

The Prompts tab in Visibility Launchpad now includes:

• A new Manage Noise toolbar button
• A per-row Mark as Noise action

When a prompt is marked as noise:

• All existing matching prompt opportunities for that website are immediately removed
• Future matching prompts are automatically ignored

Ask Widget Conversation Activity Updates

💬 Route Bad Conversations to Noise

The Ask Widget Conversation Activity screen now includes:

• A new Manage Noise toolbar button
• A per-row action that marks the conversation’s first message as noise

This makes it easy to suppress:

• spam
• garbage prompts
• accidental submissions
• repetitive low-value activity

across future visibility opportunities.

Automatic Prompt Filtering

🤖 Smarter Prompt Qualification

Ask Widget prompts are now filtered automatically before becoming prompt opportunities.

Prompts are silently ignored if they:

• are too short
• contain no meaningful words
• match customer-managed noise entries

This helps keep Visibility Launchpad focused on actionable demand signals instead of clutter.

Bulk Delete Support Added

✅ Multi-Select Deletion in Visibility Launchpad

Both the Keywords and Prompts tabs now support bulk deletion.

You can now:

• select multiple rows with checkboxes
• remove multiple items at once using Delete Selected

Additionally:

• the Delete Selected button now activates immediately when rows are checked
• no additional page interaction is required

Expanded Source Filtering

🔍 Additional Visibility Sources

The Visibility Launchpad source filter dropdown now includes:

• Eyes Over
• Search
• Ask Widget

This provides more granular control when reviewing visibility opportunities and prompt sources.

🧠 Knowledge Base Editing Fixes

The Knowledge Base edit page previously failed to render correctly when opening existing entries. Saved values such as the name, description, and URL rule patterns would not load into the form, preventing users from editing previously created entries.

This issue has now been resolved.

✅ Improvements

• Knowledge Base edit pages now render correctly for existing entries
• Saved names and descriptions load properly
• URL rule patterns are displayed and editable again
• Wildcard and regex rule configurations are now restored correctly during editing

🌐 Ask Widget Script URL Update

The Ask Widget script URL has changed.

All Ask Widget installations should now use:

https://ask.marketbrew.ai/mb-resources/js/ask-widget.js


Please update any existing Ask Widget integrations to ensure compatibility with the latest widget updates and functionality.

⚡ Flight Plan Save Performance Improvements

Saving a Flight Plan with large numbers of Ranking Sensors is now dramatically faster.

Previously, saving a Flight Plan required loading and updating each Ranking Sensor individually. As the number of selected Ranking Sensors increased, save time increased linearly — causing large Flight Plans to become progressively slower and, in some cases, risk timing out.

This process has now been optimized so save performance remains fast regardless of the number of Ranking Sensors included.

✅ Improvements

• Flight Plan save time no longer scales with the number of selected Ranking Sensors
• Large Flight Plans save significantly faster
• Flight Plans with extensive Ranking Sensor selections no longer risk stalled or delayed saves
• Improved reliability when managing enterprise-scale Flight Plans

🔎 Smarter Website Filtering Across Forecasting & Mission Control

You can now filter forecasting insights by a specific website directly from Mission Control and carry that context throughout the platform.

When a website is selected:

• The task list and forecast summary now update together
• Forecasting links preserve your selected website automatically
• The Forecasting Calculator now includes a built-in website selector with instant updates, no full page reload required

📊 Forecasting Calculator Improvements

• “Target %” renamed to Completion % for better clarity
• Added tooltips explaining completion simulations
• Removed duplicate progress indicators for a cleaner experience
• Forecast estimates are now more stable and accurate when adjusting completion percentages

✏️ Better Content Generation for Shorter Articles

Content Booster now adapts article structure more intelligently based on target word count.

Short-form articles now receive:

• Fewer, more focused sections
• Better-balanced introductions
• Shorter, more natural FAQ answers

This helps shorter content feel appropriately sized instead of overly expanded.

✨ Additional UI & Workflow Enhancements

• Website filtering moved higher within Mission Control for easier access
• Forecast summaries now clearly indicate when a website filter is active
• “SEO Performance Summary” renamed to Market Brew Impact Summary
• Forecasting workflows now maintain context more consistently between tools

Expanded Ranking Sensor coverage: Map Pack, AI Mode, and News

Ranking Sensors can now be calibrated to a broader range of search environments:

• Map Pack (Local Search): Model and measure performance in Google’s local results
• AI Mode: Calibrate sensors to AI-driven retrieval environments, including long-tail and fan-out queries
• News: Track visibility within news-focused search experiences

This expansion allows you to analyze performance across the full spectrum of how content is surfaced today—not just traditional organic rankings.

To access these new calibration options, use the manual Ranking Sensor setup wizard.

Major upgrade: Sovereign embeddings with BGE-M3

We’ve migrated our embedding infrastructure from OpenAI’s text-embedding-small model to BGE-M3, a state-of-the-art retrieval model—now fully hosted within our own infrastructure.

This is a foundational upgrade that improves accuracy, control, and performance across the entire system:

• Sovereign data control
  Your content embeddings are no longer sent to third-party APIs. All embedding generation is handled within our own environment, giving you greater control, privacy, and consistency.
• Significantly improved retrieval accuracy
  BGE-M3 is one of the top-performing models on modern retrieval benchmarks, with strong gains in:
  • Semantic search precision
  • Passage-level matching
  • Long-form content understanding
  In practice, this means:
  • Better alignment between queries and the exact passages that should be retrieved
  • Stronger clustering of your content around its true topical center
  • More reliable similarity scoring across your site and competitors
• Stronger performance in AI-driven environments
  BGE-M3 is built for modern retrieval systems, making it especially effective for:
  • AI Mode modeling
  • Fan-out query matching
  • Long-tail and low-frequency queries
• Improved multilingual and cross-domain understanding
  The model performs consistently across a wide range of topics and languages, improving stability in complex datasets.
• Ask Widget and ChatGPT Simulator improvements
  This upgrade directly enhances retrieval quality inside both the Ask Widget and the ChatGPT Simulator:
  • More precise passage selection when generating answers
  • Better grounding of responses in your actual content
  • Reduced drift and fewer irrelevant citations
  • Stronger alignment between user questions and the sections of your site that get surfaced
  The result is a more accurate, more trustworthy answer experience—closer to how modern AI systems actually retrieve and assemble responses.

This upgrade improves every feature powered by embeddings, including Ranking Sensors, Content Gap analysis, clustering, and AI Mode modeling.

Top Tasks SEO Forecasting Calculator

A new forecasting experience is now available directly from the Top Tasks page.

The Top Tasks SEO Forecasting Calculator extends the existing Ranking Sensor calculator by working across multiple Ranking Sensors at once—either within a Flight Plan or across your entire account.

• Aggregates recommended improvements across sensors
• Groups work by task type
• Estimates potential traffic and revenue impact
• Lets you model outcomes by adjusting completion levels

This provides a system-wide view of what your work can produce, not just page-by-page or query-by-query projections.

Manual traffic input for emerging and long-tail queries

During manual Ranking Sensor setup, you can now define your own traffic estimates when search volume data isn’t available.

• Enter estimated monthly volume for long-tail or emerging queries
• Or proceed without traffic modeling for those keywords

This is especially valuable when working in AI-driven environments, where many important queries don’t have traditional volume or CPC data.

Limited-data indicators on simulation tasks

Simulation tasks now clearly indicate when projections are based on limited data.

• Visual indicators highlight tasks using conservative assumptions
• Tooltips explain when data constraints may impact estimates

This gives you better context when prioritizing work and interpreting projected outcomes.

Mission Control redesign

The Mission Control overview has been redesigned to better reflect how work is organized across the system.

You’ll now see a clearer separation between:

• Site Architecture Realignment
• Answers

This makes it easier to understand where effort is being applied and how it contributes to overall performance.

✏️ Behavior Template Editing Permissions Fixed

Users who own a Behavior Template can now edit and save their changes as expected. Previously, only admins could make updates.

• Edit access now correctly applies to:
  • Template owners
  • OEM engineers managing the associated customer
• Users without permission will see:
  • Disabled instructions field and Save button
  • Tooltip explaining why editing is restricted

📊 Optimization Task Progress: Faster Insight into Gains & Losses

We’ve made it easier to track how your optimization tasks are evolving over time.

• Progress bars now include a delta badge showing change since the last recrawl:
  • 🟢 Green = improvement
  • 🔴 Red = regression
• Tasks that lose ground now display:
  • A red progress bar
  • A leading “−” indicator
• New Task History view available from snapshot listings:
  • See how task completion has changed across recalibrations

🔍 Visibility Launchpad Improvements & Fixes

Several fixes to improve accuracy and reliability within the Visibility Launchpad.

• Fixed an issue where keywords containing hyphens (e.g., “k-12 educational reform”) could not be added
• Prompts section now correctly refreshes after a site re-crawl
• Resolved stale data in gap analysis:
  • Pages returning 404s or redirects are no longer included

🛫 Flight Plan Setup Enhancements

Expanded flexibility and keyword selection during Flight Plan setup.

• New sorting option: “Top Keywords”
• Increased keyword selection per URL:
  • From 5 → 20 keywords

🧠 Content Booster — Knowledge Base Override

You can now select a specific Knowledge Base per Content Booster, overriding the site’s default set in Brand Briefing.

What’s new:

• A Knowledge Base Override dropdown now appears in the Content Booster create/edit form
  (visible when Enable Data Sources is on and at least one Knowledge Base exists)
• Select a Knowledge Base to force that Content Booster to use only that dataset
• Default option (“Use site default”) preserves existing Brand Briefing behavior
• The override is saved with the Content Booster and applied on every generation / regeneration

Why it matters:

• Gives you surgical control over context sources
• Enables testing different data sources without changing Brand Briefing
• Supports multi-strategy content workflows across the same site

API Update:

• POST /api/v1/content-booster
  → New optional field: overrideKnowledgeBaseId
• GET /api/v1/knowledge-base
  → Retrieve available Knowledge Bases and IDs

📊 Visibility Launchpad — Prompt Gap Analysis Auto-Refresh

Prompt-based gap analysis now stays fully in sync with your latest crawl data.

What was happening:

• After a crawl, keyword gap analysis updated
• But prompt-based gap records did not
• This caused stale visibility scores in the Launchpad

What’s fixed:

• Prompt gap similarity scores are now automatically recalculated after every crawl
• Results reflect the latest catalog embeddings and newly indexed content
• Visibility Launchpad now provides accurate, real-time prompt performance insights

🚀 Market Brew API Expansion — Now Available

We’re expanding the Market Brew API to give you direct programmatic access to capabilities that have already been powering results inside the platform.

Everything below reflects functionality our customers already use in the UI — now fully accessible via API so you can integrate Market Brew directly into your own systems, dashboards, and workflows.

👉 Explore and test endpoints in the API Playground:
https://brew.marketbrew.ai/openapi-ui.html
https://ask.marketbrew.ai/openapi-ui.html (Full Site Intelligence)

💬 Ask Widget API — Conversation Counting

The Ask Widget API now includes endpoints to retrieve total conversation counts across your account, with optional date range filtering.

This is the same engagement data you see in the UI, now available to:

• feed internal dashboards
• track AI adoption
• measure user interaction trends programmatically

Available via API on:
brew.marketbrew.ai
ask.marketbrew.ai (Full Site Intelligence)

✍️ Content Booster API — Counting + Advanced Listing

We’ve expanded the Content Booster API to match the flexibility of the UI.

📊 Count Endpoint
Retrieve total Content Booster entries with filters for:

• website
• title
• status

📋 Listing Endpoint (Enhanced)
The API now supports:

• filtering
• sorting
• pagination

Previously, this endpoint returned unfiltered datasets. You can now query exactly what you need, just like in the interface.

Available via API on:
brew.marketbrew.ai
ask.marketbrew.ai (Full Site Intelligence)

👂 Listen API — Keyword Monitor Management (New API Access)

The Listen system is now fully accessible via API.

Manage and analyze keyword monitoring outside the UI:

• Create, update, and delete keyword monitors
• Count or list monitors across your account or per website
• Retrieve full run history (news crawl tracking over time)
• Extract URLs and keywords from individual runs

This brings real-time monitoring workflows into your own systems.

Available via API on:
brew.marketbrew.ai
ask.marketbrew.ai (Full Site Intelligence)

📈 Top Tasks API — Optimization Intelligence (New API Access)

Top Tasks data is now available via API, enabling direct integration into reporting and execution workflows.

📋 List Endpoint
Retrieve prioritized optimization tasks:

• ranked by impact (points)
• filterable by website, Flight Plan, date range, task title
• includes sorting and pagination

📊 Metrics Endpoint
Aggregate performance insights across tasks:

• total points
• average points
• completion rate
• overall optimization progress

This is the same task intelligence used in the UI, now accessible for automation and reporting.

Available via API on:
brew.marketbrew.ai

🧠 Visibility Launchpad API — Content Gap Analysis (New API Access)

Content gap analysis from the Visibility Launchpad is now fully exposed via API.

📊 Stats Endpoint
Retrieve a high-level summary of your content gap posture:

• counts by source
• average similarity scores

🔍 Keyword Gap Endpoints

• list keyword opportunities with filtering, sorting, and pagination
• programmatically add keyword gaps with automatic deduplication

💡 Prompt Gap Endpoints

• list missing AI prompt opportunities
• add prompt gaps programmatically with deduplication

These endpoints mirror the insights available in the UI and make them actionable inside your own tooling.

Available via API on:
brew.marketbrew.ai
ask.marketbrew.ai (Full Site Intelligence)

🔗 Why This Matters

Market Brew is no longer just a platform you log into.

With this API expansion, it becomes infrastructure you can build on.

• Embed Market Brew into internal dashboards
• Power reporting systems with real ranking intelligence
• Integrate directly into CMS, workflows, and enterprise pipelines
• Automate optimization at scale

Two New Tools!

Content Summarizer

Compress any long-form content into a structured summary — bullet points, key details, and verbatim quotes — in one click.

Built on the same compression pipeline Market Brew's Content Booster uses internally when processing Knowledge Base sources before generating articles. Use it to preview exactly what the AI sees when it reads your sources, or to compress content for use in your own LLM workflows.

• Enter a URL to auto-populate the editor, or paste content directly
• Three compression levels: Low (detailed), Medium (balanced), High (concise)
• Output always includes a Detailed Summary, Key Details, and Important Quotes with speaker attribution

Embedding Separation Lab

Edit a page's content and watch its semantic similarity scores change in real time against the field of pages it currently duplicates. Lower scores mean greater separation — and less duplicate-content risk in search.

• Field Pages mode: enter a URL, edit the content in the built-in editor, click Analyze — the table shows Before %, After %, and Delta for every similar page
• Freeform mode: paste any two pieces of content and get an instant similarity score with no URL required
• Supports both page-average and top-cluster embedding comparisons, and is now linked from all new Embedding Duplication Tasks

Content Booster Improvements

WYSIWYG Editor Toolbar Always Visible
Updated the editor layout to ensure the toolbar remains visible at all times. The WYSIWYG editor is now properly sized, eliminating the need to scroll back up to access formatting controls.

“Test Content” Button Integration Fix
Fixed an issue where the “Test Content” button in Content Booster did not properly populate content and metrics in the AI Overviews Author Tool. Content now transfers correctly for immediate testing and analysis.

Removed Sticky Header Behavior
Removed the sticky behavior of the Content Booster header (blue title bar with action buttons). This resolves display issues on mobile and various screen sizes, providing a more consistent and responsive experience.

Brand Briefing — AI Generation Dialog Improvements

New Features

Crawled Content Toggle: Added a checkbox to the AI-Powered Brand Bible Generation dialog allowing you to control whether crawled website content is included in the generation. Previously, crawled content was always included. You can now uncheck "Include crawled content from [domain]" to generate the Brand Bible solely from uploaded documents.

Input Validation: The Generate button now validates that at least one content source is selected before submitting. If crawled content is disabled and no PDF documents have been uploaded, generation is blocked with a clear error message.

🐞 Bug Fixes

🔧 Knowledge Base – Save & Duplicate Handling Improved

Issue
Users could encounter errors when saving Knowledge Base entries if a full URL was entered instead of a domain. In some cases, duplicate entries were not properly detected.

Resolution
Improved input handling to support both full URLs and domain-only entries. Enhanced duplicate detection to prevent conflicting records.

Impact

• Users can now enter either full URLs or domains seamlessly
• Duplicate entries are properly identified and prevented
• More reliable Knowledge Base setup experience

🔧 Keyword Fueler – Visibility Launchpad Integration & UX Improvements

Issue
Keywords generated in Keyword Fueler were not appearing in the Visibility Launchpad, and the workflow lacked clarity.

Resolution
Fixed the synchronization issue and introduced clearer action controls by separating functionality into two buttons:

• Add to Launchpad
• Go to Launchpad

Impact

• Keywords now reliably appear in the Visibility Launchpad
• Improved user control and clarity in workflow
• More intuitive transition from keyword discovery to tracking

🔧 Content Booster – Improved Performance on Large Sites

Issue
Content Booster experienced slow startup times on large websites due to inefficient database queries when checking for embeddings.

Resolution
Optimized the embedding lookup process to eliminate unnecessary full-table operations.

Impact

• Significantly faster initialization on large datasets
• Reduced database load during startup
• Improved scalability for enterprise-level sites

⚙️ Performance Improvements

• Enhanced database efficiency for large embedding datasets
• Improved responsiveness during Content Booster initialization

‍Content Booster & Link Injector — Smarter Internal Linking

Internal links now stay on-topic

When the Content Booster or Link Injector adds internal links to your content, it previously found pages by matching the anchor phrase alone. This worked well for straightforward matches, but could send readers to unrelated pages — for example, a phrase like "snow removal" in a Bobcat article might link to a generic snow removal page rather than a Bobcat attachment page.

Links are now chosen with the full article topic in mind. The system understands what the content is about as a whole — the brand, product, and subject — and uses that understanding to prefer pages that fit both the phrase and the article's context. That means a Bobcat article will link to Bobcat pages, a skid steer article will link to skid steer content, and attachment phrases will resolve to the right attachment pages.

What this means for your content

Internal links will be more relevant and useful to readers. You'll see fewer cases where a link technically matches the anchor text but lands on a page from the wrong brand, product line, or category. Links should feel like a natural part of the article rather than an accidental detour.

Link Injector gets the same upgrade

The standalone Link Injector tool — used to add internal links to existing content — now applies the same topic-aware matching. It automatically reads your content to understand what it's about before selecting link targets, so pasted articles benefit from the same quality improvement as content generated through the Content Booster.

📊 Forecasting Calculator – Sticky Metrics

Enhancement

Metrics in the Forecasting Calculator will now remain “sticky” at the top of the screen.

Why this matters
Users analyzing projections were required to repeatedly scroll back to the top when adjusting inputs or navigating within the calculator.

Impact
• Easier to see metrics
• Reduced repetitive scrolling

🧠 Content Booster – Bio / Author Bio Template

New Template Added

A new Bio / Author Bio template is being introduced to the Content Booster library.

Why this matters
Author bios play a critical role in EEAT signals and brand authority. This template allows teams to systematically generate and optimize structured author pages.

Impact
• Standardized author bio creation
• Better authority signaling
• Expanded Content Booster coverage

🏷 Content Booster – “Mark as Published” Action

New Workflow Action

Users will be able to mark an individual Content Booster as Published.

Implementation Note
The action will be consolidated into the existing Actions dropdown for UI consistency.

Why this matters
Teams need visibility into which boosters have moved from draft to live status. This adds workflow clarity without cluttering the interface.

Impact
• Clear production tracking
• Cleaner user workflow
• Better internal coordination

📤 Top Tasks – Export to CSV

New Capability

Top Tasks data can now be exported to CSV / spreadsheet format.

Why this matters
Teams often need to share task lists externally, manipulate them in spreadsheets, or incorporate them into reporting decks.

Impact
• Downloadable task reports
• Easier client sharing
• Improved reporting flexibility

🖥 Bulk Content Booster – Scroll Fix

Bug Fix

Resolved an issue where long keyword lists in the Bulk Booster modal:
• Pushed the action button below the viewport
• Prevented scrolling inside the modal
• Caused the background page to scroll instead

Fix
The modal now properly contains its own scroll behavior and keeps action controls accessible regardless of keyword volume.

Impact
• No more blocked launch button
• Proper modal scrolling
• Improved large-scale keyword workflows

🔗 Quick Pick Screens – Crew Context Filtering

Contextual Filtering Improvement

The Link Flow Distribution Quick Pick screen will now only display sites associated with the currently selected Crew.

Why this matters
Users operating within a Crew context were seeing unrelated sites and needing to manually search for the correct one. This update aligns UI behavior with contextual expectations.

Impact
• Cleaner site selection lists
• Reduced searching
• Stronger Crew-context consistency

🛠 Optimization Task Recommendations – Heading & Meta Title Similarity Fix

Bug Fix

Optimization Task Recommendations for Heading Similarity and Meta Title Similarity were displaying "null" instead of the calculated values.

This has now been corrected.

Important:
The underlying calculations and task accuracy were not affected. This was strictly a display issue. The similarity metrics were being computed correctly in the model — they were simply not rendering properly in the UI.

Immediate Workaround (if needed)
• Click into the AI Overviews Visualizer
• Navigate to the Title / Headings tab
• The correct similarity values are visible there

For Existing Tasks
For tasks that were already generated prior to this fix:

• Flight Engineers can recalculate Tasks within the Ranking Sensor
• This will regenerate and properly display the similarity values

Impact
• Accurate similarity metrics now display correctly
• No more null values in Optimization Task Recommendations
• No impact to modeling accuracy or scoring logic
• Clear recovery path for previously generated tasks

Ask System + Ranking Sensor Performance Upgrade

This release improves AI snippet precision and significantly accelerates core modeling calculations.

1️⃣ Ask System — Accuracy & Performance Improvements

Area: Snippet Retrieval + Response Generation
Impact: Higher relevance, faster responses, stronger grounding

Overview

We upgraded the snippet retrieval layer inside the Ask system to improve passage selection accuracy and response latency.

Technical Improvements

• Refined cosine similarity scoring against page-level centroid embeddings
• Improved passage segmentation alignment for cleaner contextual boundaries
• Enhanced top-cluster filtering to reduce semantically adjacent but non-authoritative matches
• Optimized ANN vector lookup execution path
• Improved snippet de-duplication and ranking logic prior to LLM synthesis

Result

• More precise snippet selection
• Stronger alignment to authoritative pages
• Reduced hallucination risk through tighter grounding
• Faster Ask response times across large deployments

These improvements directly enhance both the Ask Widget and internal Ask workflows.

2️⃣ Ranking Sensor — 40% PSO Speed Improvement

Area: Particle Swarm Optimization (PSO) Engine
Impact: Faster modeling cycles + improved scalability

Overview

The Ranking Sensor’s PSO calculation engine has been optimized at both the algorithmic and execution layers.

Technical Improvements

• Improved fitness evaluation batching
• Reduced redundant scoring operations
• Optimized memory allocation during iteration cycles
• Improved thread scheduling and concurrency handling
• Reduced transactional overhead during boost factor evaluation

Measured Outcome

• ~40% reduction in PSO computation time
• Faster Ranking Sensor recalculations
• Increased modeling throughput across multi-site environments
• Improved scalability for larger catalogs and embedding datasets

These upgrades strengthen both the AI retrieval layer and the core modeling engine, improving real-time usability and backend performance simultaneously.

🔄 Ranking Tasks Update: Fixing Missing Tasks + Clarifying Ranking Models

We’ve released an important update to how ranking tasks are generated and displayed.

Why This Change Was Made

Users reported the following issue:

• The Top Tasks view showed no tasks
• But when navigating to the full task listing screen, tasks were visible

This created confusion — it looked like tasks existed in one place but not another.

The root cause was that two different ranking models were being used behind the scenes, and they were not aligned across screens.

This update resolves that inconsistency.

🧠 The Two Ranking Models Explained

Our ranking system uses two approaches to simulate search performance:

1️⃣ Standard Model (One Page Per Site)

This model mirrors typical search engine behavior.

• Each competing site contributes one primary page
• Rankings are built from those selected pages only
• If there are 4 competing sites, you’ll see 4 ranked pages

This keeps comparisons clean and domain-balanced.

Important limitation:
If your intended target page is not the selected “top” page for your site, tasks may not generate in this model — even if another page from your site is competitive.

2️⃣ Duplicates Model (All Pages Ranked)

This model removes the “one page per site” restriction.

• Every page from every competing site is ranked
• A single site may appear multiple times
• Rankings can include many more rows

This ensures tasks can still generate even if the target page wasn’t selected as the primary page in Standard Mode.

⚠️ What Was Happening Before

Previously:

• The Top Tasks view was hard-coded to use Standard Model only
• The Task Listing screen automatically switched to Duplicates Model if Standard produced no tasks

So in situations where:

• Standard Model produced zero tasks
• Duplicates Model did produce tasks

You would see:

• Top Tasks → empty
• Task Listing → populated

The system was silently switching models in one view but not the other.

This caused confusion and mismatched expectations.

✅ What’s Changing

🔹 No More Silent Model Switching

The system no longer automatically replaces one model with another without making it clear.

🔹 Duplicates Model Is Now Additive

Instead of “either/or” behavior:

• Standard Model remains the default
• You can optionally include results from Duplicates Model
• Both can be viewed together

This ensures tasks are not hidden simply because the target page wasn’t selected in Standard Mode.

🔹 Top Tasks Now Reflect All Available Tasks

The Top Tasks view now:

• Incorporates tasks from both models
• Removes duplicate entries
• Displays tasks that previously would have been excluded

If a task exists in either ranking model, it can now appear in Top Tasks.

🔹 Clear Labeling for Transparency

Tasks generated from the expanded (Duplicates) model are clearly identified.

This helps you understand:

• Why similar tasks may appear more than once
• Why values may differ between ranking contexts
• Which ranking logic is being applied

🎯 What This Means

• No more empty Top Tasks screens when tasks actually exist
• No more mismatched behavior between views
• Greater transparency into ranking logic
• More complete task visibility
• More reliable optimization guidance

This update makes ranking behavior consistent, predictable, and easier to understand across the entire system.

Ranking Sensor Throughput & PPC Keyword Audit Upgrades

Ranking Sensor Throughput Upgrades

Overview

We’ve introduced major throughput improvements to the calibration and calculation pipeline for Ranking Sensors. These upgrades increase parallelization across expensive calibration and scoring workloads, reducing end-to-end completion times and improving overall system throughput under load.

What Changed

• Expanded parallel execution across calibration and ranking calculation tasks.
• Reduced serialization bottlenecks in the sensor calibration pipeline.
• Improved worker scheduling to better utilize available compute resources.
• Optimized batching for high-cost calibration and recomputation stages.
• Reduced idle gaps between calibration phases and ranking recalculation phases.

Impact

• Faster completion times for Ranking Sensors performing expensive calibrations.
• Higher throughput when multiple sensors are running concurrently.
• Shorter queue times for pending sensor runs.
• Better compute utilization across processing nodes.

Behavior Notes

• No changes to calibration methodology or ranking factor math.
• No changes to scoring formulas or model interpretation.
• Results remain consistent; improvements are strictly performance-related.
• Existing Flight Plans and Ranking Sensors benefit automatically with no configuration changes required.

PPC Keyword Audit — Expanded Auditability & Usability

The PPC Keyword Audit has been enhanced to provide a stronger audit trail and more transparent support for performance-based fee models used by partner agencies.

New: Monthly Keyword & CPC Snapshots

We now store a detailed keyword-level breakdown at the time each month’s organic-value calculation is performed, including:

• Keywords used in the calculation
• GSC click counts
• CPC values used at calculation time
• Per-keyword organic value contributions

This creates a month-specific calculation snapshot that supports audit and verification of performance-based billing.

Important scope notes:

• This detailed snapshot data is available for current and future months only.
• Historical months do not have stored CPC snapshots.
• For past months, CPC values are derived using current CPC values as of today, which may differ from true historical CPC at that time.
• Audits are available only for GSC-based sources. Clients must have opted in and granted Market Brew access to their Google Search Console data.

Usability Enhancements

• Bookmarkable URLs
  Website selection and billing interval date are stored in the URL so a specific audit view can be bookmarked or shared and reopened in the same state.
• Database-level numeric filters
  Numeric columns (Clicks, Searches, CPC, Organic Value) now support minimum-value filters that execute at the database level, ensuring correct results across large datasets rather than filtering only the visible page.

Result

Partner agencies and clients now have a clearer, more defensible audit trail for organic-value calculations, along with a faster and more scalable audit interface for keyword-level inspection.

Content Booster: Published Status Feature

What's New

We've added a new Published status to help you organize your Content Booster workflow and separate finished content from work-in-progress.

Key Features

📋 Published Status

• Mark Content Boosters as "Published" when you've deployed them to your website
• Keeps your workspace clean by separating published content from drafts
• Published content is preserved and can be viewed anytime

🔘 Toggle Published Button

• Select one or more Content Boosters and click "Toggle Published"
• Each item switches between Published and its previous status
• Works with mixed selections - published items become unpublished, unpublished items become published
• Smart feedback tells you exactly what changed

👁️ Show Published Filter

• By default, published content is hidden from your Content Booster listing
• Check "Show Published" in the toolbar to include published items in your view
• Uncheck to focus only on content that still needs work

How to Use

Publishing Content

1. Select one or more completed Content Boosters
2. Click "Toggle Published" in the toolbar
3. Confirm the action
4. Selected items are marked as Published and hidden from the default view

Viewing Published Content

1. Check the "Show Published" checkbox in the toolbar
2. All content (including published) appears in the listing
3. Uncheck to return to work-in-progress view

Unpublishing Content

1. Check "Show Published" to see published items
2. Select published Content Boosters you want to edit
3. Click "Toggle Published"
4. Items revert to their previous status (Completed or Created)

Benefits

✅ Cleaner Workspace - Focus on content that needs attention
✅ Better Organization - Separate published from work-in-progress
✅ Flexible Workflow - Easily toggle status as needed
✅ No Data Loss - Published content is always accessible
✅ Bulk Operations - Toggle multiple items at once

Tips

• Use the Status filter dropdown to see only specific statuses
• The Published status appears in the status dropdown alongside other options
• Published content can still be copied, edited, or deleted if needed
• Consider marking content as Published after deploying to your live site

Need help? Contact support or check our documentation.

Top Tasks Weekly Report

We're excited to introduce Weekly Top Tasks Reports - automated email summaries that help you stay on top of your highest-impact SEO optimization opportunities.

How It Works

Every Sunday, Market Brew automatically analyzes your Flight Plans and sends personalized email reports highlighting:

• Top 10 optimization tasks ranked by total impact points
• Task occurrence counts showing how many Ranking Sensors detected each issue
• Progress tracking showing completion percentage for each task
• Smart filtering - reports only sent when Ranking Sensors have been updated with fresh data

Key Benefits

✅ Stay Informed - Never miss high-impact optimization opportunities
✅ Save Time - No need to log in to check for new tasks
✅ Prioritize Effectively - See your most impactful tasks at a glance
✅ Track Progress - Monitor completion rates week over week
✅ Mobile-Friendly - Beautiful responsive design works on any device

Who Receives Reports

Reports are sent to:

• Flight Plan owners (Crew)
• Flight Engineers
• Flight Plan team members (Team Leads and Members)

Getting Started

These notification will be enabled by default. To Disable Notifications:

1. Go to your account settings
2. Click "Notification Preferences"
3. Disable "Top Tasks Weekly Report"
4. Save your preferences

Report Details

Each report includes:

Quick Stats

• Total task occurrences across all Ranking Sensors
• Total optimization points available

Task Breakdown

• Rank (top 10 tasks)
• Task title and description
• Number of Ranking Sensors where task appears
• Total optimization points
• Current completion progress

Strategic Insights

• Guidance on template-level optimizations
• Links to view full task details in Market Brew

Smart Delivery

Reports are only sent when:

• ✅ Your Ranking Sensors have been updated with new data
• ✅ Tasks are available for the Flight Plan

This ensures you only receive reports when there's actually something new to review.

Mobile Optimized

The report email is fully responsive and works beautifully on:

• Desktop email clients (Outlook, Apple Mail, Gmail)
• Mobile devices (iPhone, Android)
• Web-based email (Gmail, Outlook.com)

Questions?

How do I stop receiving reports?
Go to Settings → Notification Preferences → Disable "Top Tasks Weekly Report"

Can I get reports more frequently?
Currently reports are sent weekly on Sundays. Contact support if you need custom frequency.

Why didn't I receive a report?
Reports are only sent when your Ranking Sensors have been updated since the last report. If no updates occurred, no report is sent.

Can I see past reports?
Log into Market Brew and visit your Flight Plan → Top Tasks to see current data anytime.

Who can I add to receive reports?
Add team members to your Flight Plan as Team Leads or Members, and they'll automatically receive reports if they enable the notification.

Centroid Keywords in Orbital Drift Stabilizer

🚀 New Feature: Centroid Keywords

We’ve added Centroid Keywords to the Orbital Drift Stabilizer to give you a fast, human-readable view of how search engines semantically categorize your content.

This feature converts embedding centroid data into clear keyword themes, helping you understand topic alignment before and during drift analysis.

Content Booster Data Sources & Research Controls

What’s New

Enabled Data Sources Toggle and Enhanced Data Source Summaries are now available in Content Booster. These upgrades give you direct control over how much external research is used during generation and significantly increase the depth and specificity of grounded references in the output.

Enabled Data Sources (Research Control)

You can now turn the research component of Content Booster on or off using the new Enabled Data Sources option in Generation Options.

What This Does

• Lets you explicitly control whether Content Booster uses:
  • Knowledge Base sources
  • Connected data sources
  • Real-time research inputs
• When enabled:
  • Content is grounded in external sources
  • Outputs include more cited facts, statistics, and verifiable details
• When disabled:
  • Generation relies primarily on your prompt and on-page context
  • Faster runs with less external enrichment

Why This Matters

Different workflows require different generation modes:

• Research ON → Best for authority content, long-form articles, and data-backed pages
• Research OFF → Best for faster drafts, rewrites, and structure-first iterations

You now have precise control over that tradeoff.

Improved Data Source Summaries

We have significantly upgraded how Content Booster reads and summarizes its research sources before writing.

What Changed

• Source articles are now summarized with:
  • Higher detail density
  • Clearer fact extraction
  • Better statistical capture
  • More concrete claims and figures
• Relevance scoring and grounded summaries are more structured and complete
• The generator receives richer research context before drafting

Result in Your Content

With these improved summaries, generated content now tends to include:

• More statistics and quantitative references
• More specific claims tied to sources
• Deeper topical coverage
• Stronger grounding and cross-reference signals
• Reduced generic filler language

Expected Impact on Generation Time

When Enabled Data Sources is ON and research is used:

• Generation may take longer than before
• This is intentional — the system is:
  • Collecting sources
  • Scoring relevance
  • Building grounded summaries
  • Cross-referencing before writing

The added time directly supports higher factual density and reduced hallucination risk.

Centroid Keywords

Centroid Keywords reveal what search engines believe your page is about, based on meaning rather than exact keyword matches.

Instead of guessing which keywords matter, you can now see the core semantic themes your content is associated with at an algorithmic level.

What This Does

Search engines don’t evaluate pages using keywords alone. They analyze meaning, context, and topical relationships.

Centroid Keywords translate those underlying signals into 8–15 clear, human-readable themes that represent how your page is understood semantically.

Think of these as your page’s semantic fingerprint.

Where You’ll See It

• Webpage Scoresheet - Found in the Key Performance Indicators section
• AI Overviews Visualizer
• AI Overviews Author‍

Why It Matters

Centroid Keywords help you:

• See your content through the search engine’s lens
  Understand how algorithms categorize your page, not just how it ranks
• Identify content gaps
  Spot missing or underdeveloped topics compared to your goals
• Improve topical alignment
  Ensure your headings, sections, and metadata reinforce the right themes
• Build stronger authority
  Create more consistent, focused content across related pages

How to Use It

1. Click Generate Centroid Keywords
2. Review the 8–15 themes identified
3. Export them using Copy Keywords as CSV
4. Compare them to your target keywords and content goals
5. Adjust content to better align with the strongest themes

This feature helps you optimize for how search engines actually understand content, not just how they match terms.

Content Booster – Improved Prompt Assistance

New Feature: “I Didn’t Like the Output”

When Content Booster doesn’t produce what you expected, you no longer need to guess how to fix it.

You can now give direct feedback and receive a custom-built focus prompt designed to correct the issue.

Faster Feedback Loop

• New “I Didn’t Like the Output” button in the Final Draft Editor
• Describe what didn’t work (tone, structure, depth, missing details, etc.)
• Receive a revised focus prompt tailored to your feedback

Smarter Prompt Guidance

• Analyzes your feedback and current content settings
• Generates a clearer, more precise focus prompt
• Preserves intent while tightening direction
• Designed to be copied directly into the Content Strategy tab

Clear, Simple Workflow

1. Click I Didn’t Like the Output
2. Explain what needs improvement
3. Copy the improved focus prompt
4. Paste it into Content Strategy
5. Regenerate content with better guidance

Why This Matters

Content quality issues are often prompt clarity issues, not writing issues.

This feature helps you:

• Turn vague frustration into actionable direction
• Gain confidence in what the prompt is actually instructing
• Refine content without trial-and-error guessing
• Achieve better results faster with fewer regenerations

It’s ideal when a keyword alone isn’t enough, but writing a perfect prompt feels unclear.

Content Booster – New Content Type: Press Release

Press Release is now available as a dedicated content type in Content Booster.

This option is designed specifically for announcements, launches, partnerships, funding news, and other official statements that need a clear, structured, and authoritative format.

Why We Added This

Press releases follow a very different structure than blogs or guides. They prioritize clarity, attribution, factual accuracy, and scannability over long-form exploration.

Using a generic blog format for press releases often leads to:

• Overly narrative introductions
• Missing or unclear announcements
• Weak headlines and summaries
• Inconsistent formatting for media use

This new content type removes that friction by guiding Content Booster toward a press-ready structure from the start.

What’s Different About Press Release Mode

When Press Release is selected, Content Booster will:

• Emphasize a clear, news-focused headline
• Lead with the announcement, not background context
• Use a formal, neutral, media-friendly tone
• Structure content for quick scanning by journalists and readers
• Focus on facts, attribution, and relevance over opinion

The goal is content that can be published directly or shared with media without heavy rewriting.

Where to Find It

• Open Advanced Settings
• Expand Content Settings
• Select Press Release from the Content Type dropdown

When to Use It

Use Press Release when you are:

• Announcing a product launch or feature
• Sharing company news or updates
• Communicating partnerships or milestones
• Publishing official statements intended for external distribution

For thought leadership, education, or exploration, Blog Post or How-To formats remain the better choice.

Content Booster Update: URL Support in Focus Prompts

Feature: Smart URL Summaries in Focus Prompts

Content Booster now supports URLs inside Focus Prompts.

When you include a link in your Focus Prompt, the system automatically detects it and gives you the option to replace the URL with a concise summary of its content.

Instead of passing a raw link to the AI, Content Booster converts the page into a clear, readable summary that becomes part of the prompt itself.

Why We Designed It This Way

Raw URLs hide important details.

If an AI is given a link, there’s no way for you to know which parts of that page are being used, what context is being pulled, or how that information is influencing the output.

That’s why Content Booster does not allow URLs to remain in Focus Prompts.
You must replace them with a summary.

This ensures you always know exactly what the prompt is doing.

You can read the summary, edit it if needed, and see—line by line—what information is guiding content generation. Nothing is implicit or hidden.

How It Works

1. Paste a URL into your Focus Prompt
2. Content Booster detects the URL automatically
3. Click “Summarize URLs in Focus Prompt”
4. The URL is replaced with a concise, structured summary of that page

That summary is what the AI uses—no live fetching, no invisible interpretation, no guesswork.

Why This Matters

• Full transparency – You can see exactly what context is being applied
• Predictable results – Prompts behave consistently because inputs are explicit
• Better grounding – Content is anchored to known facts, not inferred ones
• Lower risk of hallucinations – The AI works from vetted summaries, not opaque links

When to Use This

This feature is ideal when:

• You want to reference authoritative sources with precision
• Accuracy and traceability matter
• You need to align content with a specific historical, legal, or editorial source
• You want confidence that the AI is responding to your intent—not interpreting a link on its own

Summary

URLs are now supported—but only as something you can fully see and control.

By forcing URLs to become readable summaries, Content Booster ensures that you always understand how your Focus Prompt is shaping the output. What you see in the prompt is exactly what the system is using.

📚 New Feature: Knowledge Bases for Content Generation

We’ve introduced Knowledge Bases, a powerful new way to control where your AI-generated content gets its information. This gives you more accuracy, more trust, and more consistency in every piece of content you create.

🌟 What Is a Knowledge Base?

A Knowledge Base lets you define a set of trusted websites that Market Brew should use when researching and citing sources for content.

Instead of pulling information from the open web, your content can now be grounded in sources you choose—such as industry publications, research sites, or your own properties.

✅ Why This Matters

• ✔ More accurate and reliable content
• ✔ Citations from sources you trust
• ✔ Better alignment with your brand, industry, or mission
• ✔ Less risk of low-quality or irrelevant references

🧠 How It Works (At a High Level)

1. You create a Knowledge Base
2. You add trusted websites (and optionally limit which sections are used)
3. You assign that Knowledge Base to your website via the Brand Bible
4. Content generation automatically uses those sources for research and citations

No extra steps during content creation—everything happens automatically once configured.

🔧 What You Can Do Now

1️⃣ Create Knowledge Bases

• Organize sources by topic, industry, or purpose
• Reuse the same Knowledge Base across multiple content projects
• Add descriptions so your team knows how each one should be used

2️⃣ Choose Exactly What Gets Indexed

For each website you add, you can decide:

• Entire Site
  Use everything from a trusted source
• Include Only Certain Sections
  Example: only /blog/ or /research/
• Exclude Certain Pages
  Example: remove job listings or outdated content
• Specific Pages Only
  Perfect for curated articles, research papers, or official documents

This gives you full control without needing technical expertise.

3️⃣ Assign a Knowledge Base to Your Brand Bible

Once a Brand Bible exists, you can:

• Select a Knowledge Base from a dropdown
• Save it once
• Automatically apply it to all future content generation for that site

✍️ Smarter Citations & Links (Automatically)

When a Knowledge Base is used:

• Content cites information from your selected sources
• Citations are clearly numbered and linked
• External links respect your link-density and ratio settings
• If the link limit is reached, extra citations are still shown (just not linked)

This keeps your content clean, credible, and SEO-friendly.

🔄 Seamless Upgrade from the Old Friend/Foe System

If you were already using Friends:

• They’ve been automatically converted into Knowledge Bases
• Each trusted site was preserved
• No action required—everything keeps working

(Previously blocked “foe” sites are not included, by design.)

🖥️ New Screens You’ll See

• Knowledge Base Management
  • Create, edit, and organize your Knowledge Bases
• Data Source Editor
  • Add websites
  • Control which pages are included
  • Easily adjust rules visually
• Brand Bible Integration
  • Select which Knowledge Base powers your content

Everything is fully integrated into your existing workflow.

🧭 Best Practices

• Create separate Knowledge Bases for different content goals
  (e.g., “News Sources”, “Research”, “Internal Docs”)
• Start broad with trusted sites, then narrow if needed
• Review your sources occasionally to keep content fresh and relevant

🚀 What This Unlocks

Knowledge Bases are the foundation for future enhancements like:

• Source performance insights
• Shared team Knowledge Bases
• Smarter citation scoring
• Advanced content validation

This release sets the stage for even greater control and transparency in AI-powered content creation.

If you have questions or want help setting up your first Knowledge Base, our support team is ready to help.

Happy publishing!

🔍 Ask: Smarter Conversation Search

Finding past conversations in Ask is now fast and effortless.

What’s New

• Search your conversations by meaning, not just titles
  You can now find past Ask conversations based on what was actually discussed, even if you do not remember the exact wording.
• Automatic search when you need it
  When you have a larger conversation history, Ask intelligently adds a search option so older conversations are never lost.
• Real-time results
  As you type, relevant conversations appear instantly, ranked by relevance.

Why It Matters

• Quickly recover insights from older conversations
• No more scrolling through long conversation lists
• Finds conversations based on intent, not exact phrasing

📰 Content Booster: Smarter Use of Current Events

Content Booster now knows when your content needs up-to-date information and proves it.

What’s New

• Automatic detection of time-sensitive sections
  Content Booster evaluates each section to determine whether current events are required.
• Relevant news sources added only when needed
  When freshness matters, Content Booster pulls in the most relevant news context to improve accuracy and credibility.
• Clear source transparency
  You can now see exactly:
  • Which sections used news sources
  • Why current events were included
  • Which sources were referenced
• Built-in reporting
  A new “News Sources Used” section appears in the final draft when applicable, making content review and validation easy.

Why It Matters

• Prevents outdated or inaccurate content
• Avoids unnecessary news clutter in timeless content
• Builds trust with editors, stakeholders, and readers
• Makes AI-generated content easier to verify and defend

✨ Overall Impact

These updates continue Market Brew’s shift toward:

• More discoverable insights
• More trustworthy AI-assisted content
• More clarity around why decisions are made

Ask helps you find knowledge faster.
Content Booster helps you publish with confidence.

✨ AI Recommendations Now Available in the AI Overviews Visualizer

We have added AI Recommendations directly into the AI Overviews Visualizer and Author, making it easier to move from analysis to action without leaving your workflow.

What’s New

🧠 AI Recommendations Button
A new AI Recommendations button is now available on the Content Similarity tab. This provides guided rewrite suggestions tailored to how embeddings are being interpreted by AI systems.

🔍 Embedding-Level Guidance
Clicking on any individual embedding now opens a right-hand sidebar with AI Recommendations on how to rewrite or improve that specific passage. This experience mirrors what users already rely on in the Orbital Drift Stabilizer, bringing consistency across tools.

Why This Matters

Instead of guessing how to improve similarity or alignment, you can now see exactly how Market Brew recommends rewriting content at the passage level, based on the same modeling used throughout the platform.

This closes the gap between:

• Seeing a similarity issue
• Understanding why it exists
• Knowing what to change

What’s Coming Next

This release is part one of a broader initiative to unify Market Brew’s content analysis and writing tools into a single, cohesive experience. Future enhancements will continue to reduce friction between insight, recommendation, and execution as we build toward a fully integrated content analysis and writing suite.

More to come.

‍Clearer Signals. Faster Content. Smarter Decisions.

This release focuses on improving signal accuracy, removing confusing UX states, and making content creation faster and more intuitive. The goal is simple: help you trust what you see, act faster, and stay aligned with how search and AI systems actually interpret your site.

🧭 Visibility Launchpad: Most Similar Webpage now links to AI Overviews Visualizer

The Most Similar Webpage link in Visibility Launchpad now opens the AI Overviews Visualizer instead of the Webpage Scoresheet.

What this means for you

• The link now reflects how the match is actually generated
• The page shown is the exact URL reported by the rankings data (from the Mine tab)
• The AI Overviews Visualizer is used because this match is based on a fast embedding pass, not a full crawl
• You no longer see conflicting messages about whether a page has been crawled or processed

This creates a clearer, more honest experience when reviewing similarity matches.

📘 Brand Bible "Not Specified" Fix

Brand Bible fields that previously displayed “not specified” across the platform have been corrected.

What this means for you

• Your saved brand guidance now displays properly
• Prompts and generated content reflect your real inputs
• No more placeholder values appearing in workflows

✨ Content Booster: Faster Defaults, Smarter Prompts

Two major improvements make Content Booster more fluid and more aligned with your goals.

Auto-Expand Focus Prompt

When your prompt is short or vague, Market Brew now automatically expands it using your brand, page intent, and surrounding context.

You get

• Richer, more specific content
• Less manual prompt engineering
• Better alignment to real search intent

Optional Advanced Settings

The step-by-step wizard has been simplified.

You can now

• Write a prompt and click Ignite Content immediately
• Open advanced controls only when you want them
• Stay in flow without mandatory setup steps

This creates a fast default path with power controls still available when needed.

‍📈 Historical Drift

Historical drift is now available in the Orbital Drift Website Visualizer.

You can now see how each website’s alignment changes over time, with Max / Avg / Min drift values plotted across the timeline.

What this means for you

• Track when a site starts drifting from its original intent
• See whether optimizations are holding or slipping
• Compare stability across time instead of relying on single snapshots

Summary

This release sharpens Market Brew’s signal integrity, simplifies content creation, and removes sources of confusion that slow teams down.

The result

• Clearer insights
• Faster workflows
• More trustworthy signals
• Stronger alignment between strategy and execution

SEO moves from reacting to noise to acting on outcomes.

New Tool: Link Injector
Automated internal link deployment using semantic matching

We’ve released Link Injector, a new tool that analyzes your content and injects contextually relevant links to pages within a selected website using Market Brew’s semantic similarity engine.

It completes the internal linking workflow:

Spot → Find → Inject
Discovery → Targeting → Deployment

Where to find it
Tools → Link Analysis → Link Injector

Works with any licensed website in your account that has embeddings indexed.

What it does
• Finds linkable phrases in your content
• Matches them to relevant pages using embeddings
• Injects links automatically
• Preserves your existing HTML and formatting

Key controls

Similarity Threshold (30%–85%)
Controls how closely a phrase must match a page before a link is created.
Default: 40% (Relaxed)

• 30–40% → very broad matches
• 40–55% → relaxed (default)
• 55–70% → balanced
• 70–85% → strict

Link Density
Set how many links to add per 100 words (1.0–5.0).
This controls how aggressively authority is distributed.

Target Website
Choose which licensed site will receive the links.

‍Common uses
• Updating older content
• Preparing AI content for publishing
• Migrating content
• Standardizing internal linking

Link Injector turns content into a modeled authority distribution engine — not guesswork, not manual linking, but repeatable deployment based on semantic relevance.

Content Booster: Current Events Intelligence - Release Notes

Feature: Automatic Current Events Detection & Integration

🆕 What's New

Content Booster now automatically detects when your content topics benefit from current events and seamlessly integrates relevant, real-time information into your generated articles.

Intelligent Current Events Detection

When generating content, Market Brew now:

• Analyzes each section to determine if current events would enhance the content
• Automatically searches news sources when topics like "latest trends," "recent developments," or "current state" are detected
• Retrieves semantically relevant snippets from top news articles and stories
• Integrates information naturally without forcing irrelevant current events into evergreen content

✨ Key Benefits

1. Always-Current Content

Your generated articles automatically include the latest industry developments, market trends, and breaking news when relevant to your topic.

2. Zero Manual Research

No need to manually search for current statistics, recent case studies, or latest industry news—Market Brew does it automatically during content generation.

3. Smart Detection

The system intelligently distinguishes between:

• Evergreen topics (how-to guides, foundational concepts) that don't need current events
• Time-sensitive topics (market analysis, technology trends, policy updates) that benefit from real-time information

4. Semantic Relevance

Using advanced embedding technology, Market Brew finds the most contextually relevant news snippets—not just keyword matches—ensuring current events genuinely support your content.

🎯 When Current Events Are Added

Current events are automatically detected and added for sections involving:

✅ Topics that benefit from current events:

• Recent industry trends or developments
• Current market conditions or statistics
• Latest technology advancements
• Policy changes or regulatory updates
• Competitive landscape analysis
• State-of-the-art practices or best practices that evolve
• Topics with explicit time references ("recent," "latest," "current," "2025," "2026")

❌ Topics that remain evergreen:

• Foundational how-to content
• Historical facts or completed events
• Timeless principles and frameworks
• General definitions and explanations
• Theoretical or conceptual topics

🔧 How It Works

Behind the Scenes

1. Section Analysis: Each content section is analyzed to determine if current events would add value
2. Confidence Scoring: The system generates a confidence score (0-100%) for whether current events are needed
3. Smart Search: If confidence exceeds 65%, Market Brew generates an optimized search query
4. News Retrieval: Relevant news articles and top stories are crawled and processed
5. Semantic Matching: Using embeddings, the most relevant snippets are identified based on your section's context
6. Natural Integration: Current events are provided as context to the content generator, which integrates them naturally when appropriate

What You'll Notice

• Sections with current data: When Market Brew adds current events, you'll see references to recent developments, latest statistics, or breaking industry news
• Source diversity: Information comes from authoritative news sources, industry publications, and top stories
• Natural flow: Current events are woven into the narrative, not awkwardly inserted
• Graceful degradation: If news retrieval fails, content generation continues normally without current events

📝 Example Impact

Before (without current events):

"SEO strategies continue to evolve as search engines improve their algorithms. Businesses should focus on quality content and user experience."

After (with current events):

"SEO strategies continue to evolve as search engines improve their algorithms. Recent analysis from January 2026 shows that Google's latest core update has increased emphasis on E-E-A-T signals, with sites demonstrating clear expertise seeing ranking improvements of 15-30%. Businesses should focus on quality content and user experience while ensuring proper author attribution and topical authority."

Content Booster Release Notes - FAQ Enhancement Update

New Feature: Intelligent FAQ Generation with Current Events

We've significantly enhanced how Content Booster handles FAQ sections in generated content, making your FAQs more accurate, relevant, and up-to-date.

Smart FAQ Question Detection

• FAQ sections now use the actual questions defined in your content strategy (Stage 1)
• Each FAQ question is analyzed individually to determine if current events would improve the answer
• Questions about recent developments, current challenges, or evolving situations automatically trigger news retrieval

Current Events Integration for FAQs

• FAQ answers are enriched with recent news and developments when relevant
• The system intelligently determines which questions benefit from current information
• Multiple layers of analysis ensure comprehensive coverage:
  • Overall FAQ section analysis for broad context
  • Individual question analysis for specific current events needs
  • Smart confidence thresholds prevent irrelevant news from cluttering answers

Natural, Accurate Responses

• Current events are integrated naturally into FAQ answers
• No fabricated statistics or invented case studies
• Clear guidelines prevent AI buzzwords and jargon
• Answers remain focused on providing genuine value to readers

How It Works

1. Question Analysis: When generating an FAQ section, Content Booster analyzes each question to detect if it relates to current events, recent developments, or evolving situations
2. Intelligent News Retrieval: Questions that benefit from current information trigger targeted news searches:
   • Broad searches for questions covering multiple recent topics
   • Focused searches for specific current event questions
   • Balanced snippet collection (4-8 articles) to avoid information overload
3. Contextual Integration: Current events are woven naturally into answers:
   • Recent developments support your answers with real data
   • Up-to-date information keeps your content relevant
   • Questions that don't need current events remain focused on core concepts

Benefits

• More Accurate FAQs: Answers reflect current reality, not outdated information
• Better User Experience: Readers get relevant, timely information
• Improved SEO Value: Fresh, current content performs better in search results
• Reduced Manual Updates: Current events are automatically incorporated at generation time

‍
Website Crawl 403s - Crawler Enhancement
Previously when a site was crawled and it returned 403 errors, we only enabled a proxy crawl when we had crawled at least 50 pages from the site. This causes many sites to never get included in the Ranking Sensors b/c we didn't reach this limit. Now all sites that exceed a 403 ratio will be automatically switched to proxy crawling, resulting in these sites included in many more Ranking Sensors, resulting a higher PSO score for all Ranking Sensors.

This release improves workflow clarity, historical insight, ranking accuracy, and operational control across core platform surfaces.

📚 Brand Bible Enhancements

Historical Brand Bible Listing

A new Brand Bible Listing screen now shows historical versions of each Brand Bible entry.

Why this matters

• Track how your brand narrative evolves over time
• Compare previous guidance to current positioning
• Maintain consistency across AI-generated content, Ask, and self-healing workflows

📊 Mission Control & Task Intelligence

Top Tasks Now Reflect Selected Website

When a website is selected in Mission Control, the Top Tasks panel now dynamically reflects tasks for that specific site.

Benefits

• Immediate, site-specific prioritization
• Clearer operational focus in multi-site environments
• Reduced noise from irrelevant tasks

Target Website Filter in Top Tasks View

The Top Tasks screen now includes a Target Website filter, allowing you to explicitly scope tasks to a specific site.

Benefits

• Rapid context switching between websites
• Cleaner task views for multi-domain customers
• Improved focus for execution teams

📈 Ranking & Sensor Accuracy

Manual Ranking Sensor Review Fix

Resolved an issue where the Manual Ranking Sensor Setup Wizard displayed “Not Ranking” on the review screen, even when the site was ranking.

Result

• Accurate ranking validation during setup
• Greater confidence before launching calibrated sensors

🚀 Flight Plan Management

Copy / Delete Actions Restored

Fixed an issue where Flight Plans could not be copied or deleted from the listing view.

Outcome

• Full lifecycle control over Flight Plans
• Faster iteration when testing SEO strategies

Summary

This release tightens the feedback loop between strategy, execution, and measurement by:

• Making Brand intelligence auditable over time
• Aligning Top Tasks with live operational context
• Enabling site-level task filtering
• Ensuring ranking signals are trustworthy
• Restoring full control over Flight Plan workflows

These improvements continue to move Market Brew toward a fully self-healing, decision-driven SEO system.

Bug Fixes

Listen Usage Limit Display Fix

• Fixed an issue where the Listen module displayed “0 / 15” while also indicating that the maximum limit had been reached.
• Usage limits are now correctly enforced and displayed per crew, rather than per customer, ensuring accurate reporting and preventing false limit exhaustion.

Website Embedding Domain Cluster Index Fix

• Resolved an issue where the domain cluster value could appear as -1 in the Website Embedding Listing.
• Domain cluster identifiers now correctly reflect valid clustering assignments.

Enhancements & Feature Updates

Improved Search Interaction for Website Embedding Listing

• Updated the Website Embedding Listing to use a dialog-based prompt input instead of an inline search field, providing a cleaner and more intentional search experience.
• Added direct linking from the Brand Briefing screen to the Website Embedding Listing for improved workflow continuity.

Expanded Keyword Visibility for Embedding Clusters

• Increased the number of displayed centroid keywords and branded keywords from 3 to 12, offering deeper visibility into the dominant semantic signals driving cluster classification.

AI Overviews Author & Orbital Drift Stabilizer – Meta Information Control

• Added a new “Add Meta Information” button, allowing users to explicitly include and manage metadata as part of AI Overviews alignment and orbital drift stabilization workflows.

Core Content Sample (New Feature)

• Introduced the Core Content Sample across the Website Dashboard and Website Embedding Listing screens.
• The Core Content Sample is based on approximately 15 representative embeddings from the largest content cluster (Domain Cluster #1).
• This provides a concise, high-signal snapshot of how embedding-based systems interpret the site’s primary topical focus, analogous to cluster sampling used in AI Overviews and passage-based retrieval.

Content Booster – Recent Enhancements

New: Article Layout Configuration

A new Article Layout section has been added to the Content Booster setup screen, giving users finer control over how generated content is structured.

Key capabilities include:

• Generate Introduction / Abstract
  Choose whether to automatically generate an introductory or abstract section at the start of the content.
• Custom Layout Instructions
  Define high-level structural rules that guide how the content is organized (e.g., narrative flow, step-based structure, verse-by-verse analysis).
• Section Heading Options
  Choose whether to include visible section headings or allow content to flow continuously as prose.
  When headings are disabled, logical section breaks are still respected internally, enabling narrative-style content without visual headers.

This enhancement enables significantly more flexibility for long-form, narrative, or editorial content while maintaining consistent internal structure.

Enhanced: “Use Site Context” (Internal Linking Intelligence)

The Use Site Context option has been substantially upgraded when Internal Links are enabled.

What’s changed:

• Content Booster now attempts to incorporate additional supporting content from existing site pages, based on the entire generated section, not just anchor text.
• This allows the system to surface more relevant references and internal links that better reflect topical alignment and on-site authority.

How this improves results:

• Previously, the system identified anchor text first and then searched for the closest snippet embeddings to determine internal link targets.
• The upgraded approach evaluates full section context, resulting in more meaningful internal references and better semantic alignment.

This behavior now mirrors the proven logic used in Friends/Foes, where defining a friend site and enabling external links allows Content Booster to add contextual quotes or reference links automatically.

Improved: Stage 6 – Editorial Rules Enforcement

Stage 6 (Final Output) has been enhanced to more closely adhere to the Brand Bible’s editorial_rules section.

Impact:

• Tone, style, formatting, and editorial constraints defined in the Brand Bible are now enforced more consistently.
• Final output more accurately reflects brand voice, editorial standards, and compliance requirements without additional manual editing.

This results in content that is both search-aligned and brand-aligned at publication time.

New: Reduced Minimum Word Count

The minimum word count for Content Booster output has been lowered to 500 words.

Why this matters:

• Shorter-form content now works more effectively with custom section layouts and no section title configurations.
• Enables narrative, editorial, and commentary-style content that benefits from tighter structure without unnecessary length.
• Provides greater flexibility when producing supporting articles, topical explainers, or focused authority pieces.

This change was validated alongside the new Article Layout controls and is designed to expand Content Booster’s use cases without sacrificing semantic depth or internal structure.

Summary

These updates collectively improve:

• Structural control over generated content
• Internal linking accuracy and contextual relevance
• Editorial consistency and brand compliance
• Flexibility for both long-form and concise narrative content

Together, they move Content Booster closer to a fully guided, first-principles content generation workflow that aligns structure, semantics, and brand standards in a single pass.