Brand Stack Changelog

Sign-in polish, dark mode fix & workspace security hardening

Sun, 17 May 2026 12:00:00 GMT

Sign In and Sign Up are now accessible from every public page on brandstack.dev. A first-session dark mode bug that caused the sign-in modal to flash white then crash is fixed. Workspace publish actions are now strictly scoped so content from one workspace can never be touched by another.

FIXES: Sign in from any page, not just the home page
When browsing public Brand Stack pages like the community gallery, pricing, docs, or changelog, visitors can now sign in or create an account directly from the left sidebar. Previously, only the home page had sign-in options in the navigation — visitors who discovered Brand Stack through a public persona or tool had no visible path to get started. Sign In and Sign Up buttons now appear in the sidebar on every public page when you're not logged in. They disappear once you're signed in, replaced by the existing "Open workspace" link.

FIXES: Sign-in modal no longer flashes light mode in dark mode
If you had Brand Stack set to dark mode, the first time you opened the sign-in or sign-up modal in a browser session it would briefly render in light mode, then crash with a "Something went wrong" error when you tried to close it. A second attempt always worked — the bug only hit on the very first open. The root cause was a timing issue where the dark theme loaded a few hundred milliseconds after the modal appeared, then Clerk couldn't cleanly handle the swap mid-flight. The fix loads the dark theme before the modal can open, so it renders correctly on the first try.

FIXES: Your workspace content is locked to your workspace
All publish and unpublish actions across products, projects, documents, and scenes now verify that the content belongs to your workspace before making any change. Previously, a gap in the authorization logic meant that under certain conditions, a request could affect a resource from a different workspace if the resource ID was known. Every publish action now applies a workspace ownership check at both the read and write steps. If the content doesn't belong to your workspace, the action is rejected with a clear error. Your data stays yours.

Free Tools, Moonshot Test, Studio Redesign & Usage Dashboard

Fri, 15 May 2026 12:00:00 GMT

Brand Stack ships a free tool library for founders and marketers — including The Moonshot Test for AI-powered idea validation, a Color Palette extractor, UTM Builder, and OG Image Preview & Builder. Plus: a full credit usage ledger, a cleaner Mockup Studio prompt, smarter document saving, and admin-only billing controls.

NEW: The Moonshot Test — Free AI Idea Validator
Is your startup idea a vitamin or a painkiller? **The Moonshot Test** is a free tool at [brandstack.dev/tools/moonshot-test](/tools/moonshot-test) that vets your idea against six classic startup frameworks — including the Toothbrush Test, the 10× Test, and the Contrarian Truth — and returns a full scorecard with letter grades. More than a grade, it gives you a **diagnosis** of your idea's specific strengths and weaknesses, three **Elevation Levers** you can push to make it bigger, and **three progressively bolder reframings** of your idea — with the third one being the full moonshot. The tool is designed to challenge rather than flatter. No account required. After you run the test, you can **share your result** to a public community gallery where other founders can discover it — each shared result gets its own indexed page at a unique URL, so you can send it to investors or collaborators. [Try the Moonshot Test →](/tools/moonshot-test) · [Read the docs →](/docs/tools/moonshot-test)

NEW: Free Tools Library at brandstack.dev/tools
Brand Stack now has a public **free tools library** at [brandstack.dev/tools](/tools) — a growing catalog of no-signup, no-ads tools for the jobs founders and marketers do every week. Three new tools launched alongside the catalog: **[OG Image Preview & Builder](/tools/og-image-preview)** — Paste any URL to see how its social card renders on X, LinkedIn, Facebook, iMessage, Slack, and Discord. Spot missing tags, fix warnings, then switch to Builder mode to design a 1200×630 OG image from a template — customize colors, logo, and text, then export as WebP or PNG. As you design, a live preview shows exactly how the finished image will look on each network. **[UTM Builder](/tools/utm-builder)** — Build clean, validated campaign URLs for Google Analytics. Ten quick-start channel templates (Newsletter, LinkedIn, Google Ads, etc.) prefill source and medium. Validation catches the three issues that silently break GA reports: spaces, uppercase letters, and mixed casing. Previously used values are saved as autocomplete suggestions, and you can share a pre-filled form link with teammates. **[Color Palette from Image](/tools/color-palette-from-image)** — Drop, paste (⌘V), or upload an image to extract its dominant color palette. Choose between 3, 5, 7, or 9 colors. Each swatch shows HEX, RGB, and OKLCH values, the percentage of pixels it covers, and WCAG AA contrast scores against white and black — so you know immediately which colors are accessible as backgrounds. Runs entirely in your browser; nothing is uploaded. [Browse all free tools →](/tools) · [Read the docs →](/docs/tools)

NEW: Credit Usage Dashboard
**Settings → Usage** is a new tab available to all workspace members. It shows your workspace's complete credit history as a bank-statement-style ledger — every task consumed, credit pack purchased, subscription renewal, refund, free grant, and expiration in one chronological view. Summary cards at the top show your current balance, period usage, lifetime usage, and an estimated lifetime spend. Each task-consumption row links directly to the document, persona, or mockup the credits were spent on, so you can trace exactly where your credits went. [Read the docs →](/docs/billing/usage-dashboard)

FEATURE: Smarter Document Saving
Documents now save with **⌘S / Ctrl+S** or the **Save** button in the editor toolbar. A save-status indicator in the toolbar shows when your document was last saved and flags unsaved changes. Brand Stack also auto-saves every 10 seconds whenever there are pending changes, so nothing is lost if you close a tab or navigate away. Previously, documents used a real-time sync model. The new explicit save model is simpler, faster, and gives you clear control over when content is committed. [Read the docs →](/docs/documents/research-and-collaboration)

FEATURE: Mockups & Scenes Together on Persona Pages
Persona detail pages now show **both mockups and scenes** in a single unified gallery. Previously, only mockups appeared. Now you can see all generated visuals for a persona in one place, launch new scene or mockup generations from the same strip, and delete or unlink either type without jumping to another view.

FEATURE: Cleaner Mockup Studio Prompt
The Mockup Studio prompt input has been redesigned with a **two-row layout** that keeps the text input and Generate button together, and puts attachment actions — screenshot, persona, and reference image — on a dedicated row below. Each attachment slot is always visible as a named placeholder, so you can drop images or open pickers directly from the input without hunting through menus. File size limits (5 MB) are now enforced up front with a clear message before you try to generate.

FEATURE: Admin-Only Billing Controls
Workspace **admins** are now the only members who can upgrade plans, purchase credit packs, invite members, or manage roles. Non-admin members can still view all subscription and billing information — they simply can't make changes. Disabled actions show a tooltip explaining why, and a banner on subscription pages confirms the restriction. This prevents accidental billing changes from non-admin team members in shared workspaces.

MCP Connector — Brand Stack in Claude Code, Cursor, Codex, and Claude.ai

Sun, 10 May 2026 12:00:00 GMT

Connect Brand Stack to any MCP-compatible AI client with a single OAuth handshake. Read your brands, products, personas, projects, documents, and assets — and generate mockups and scenes — directly from a chat prompt. No keys to paste, no SDK to install.

NEW: The Brand Stack MCP connector
Brand Stack now ships with a full [Model Context Protocol](https://modelcontextprotocol.io) server at `https://app.brandstack.dev/api/mcp`. Add it once to **Claude Code**, **Claude.ai web**, **Cursor**, **Codex**, or any MCP-compatible client and your AI tool can read your workspace and create mockups and scenes from natural-language prompts. The connector ships with 13 tools — 11 read tools (brands, products, personas, projects, documents, assets, credits) plus two write tools (`create_mockup`, `create_scene`) that consume credits exactly like the in-app studio. The model picks the right tool and fills in the arguments; you just describe what you want. [Read the docs →](/docs/mcp) · [Try the marketing page →](/mcp) · [Browse example prompts →](/docs/mcp/prompts)

NEW: One-click OAuth, no API keys
Adding the connector kicks off a browser-based OAuth 2.1 handshake. Sign in with your existing Brand Stack account, pick which workspace the client gets access to (admin-only), click **Allow**. The client receives a short-lived access token plus a 30-day rotating refresh token — nothing for you to copy, paste, or store. Revoke at any time from **Settings → Developer**. ``` claude mcp add --transport http brandstack https://app.brandstack.dev/api/mcp ``` The same URL works in Claude.ai web (**Settings → Connectors → Add custom connector**) and as an entry in your Cursor or Codex `mcp.json` / `config.toml`.

NEW: Generate mockups and scenes from a chat prompt
The new `create_mockup` and `create_scene` MCP tools wrap the same generation pipeline as the in-app studio. From Claude or Cursor: > "Create a 16:9 phone mockup featuring my persona Sarah, with natural lighting and an over-the-shoulder camera angle." The agent calls `list_personas_for_product` to find Sarah's id, then `create_mockup` with the right parameters. Credits are charged once at the start and refunded automatically on any failure path — including a brand-new periodic sweeper that catches rows stuck in `generating` if the underlying request was killed mid-pipeline. Multi-step workflows work too: > "Pick the most senior persona from my Acme product and create three scene variations of them in different environments."

NEW: Developer settings, rebuilt around MCP
The **Settings → Developer** page is now MCP-first. Copy-paste install commands for Claude Code, Claude.ai, Cursor, and Codex sit at the top. Below that, a "Try asking your AI" list of copyable prompts so you know what to type once connected. Authorized clients show up as a revocable list — one click kills every active token for that grant.

NEW: Public marketing page at brandstack.dev/mcp
A new public-facing page at [brandstack.dev/mcp](/mcp) explains what the connector does, how to install it in each client, and shows eight copyable example prompts. Linked from the docs nav and the Developer settings page.

FEATURE: Docs reorganized around the connector
A new **MCP Connector** section in the docs covers everything end-to-end: - [Overview](/docs/mcp) — how the OAuth handshake works, what tokens and scopes look like. - [Setup](/docs/mcp/setup) — client-by-client install for Claude Code, Claude.ai, Cursor, Codex. - [Example prompts](/docs/mcp/prompts) — copyable prompts grouped by capability. - [Tools reference](/docs/mcp/tools) — every tool with inputs, outputs, and example prompts. - [Troubleshooting](/docs/mcp/troubleshooting) — fixes for OAuth flow, scope, and tool errors. The previous `/docs/api` page covering the REST v1 API has been removed — see "FIXES" below.

FEATURE: Built on the open standard
Brand Stack implements MCP 2025-06-18 with OAuth 2.1 + PKCE (S256, mandatory), Dynamic Client Registration (RFC 7591), protected-resource metadata (RFC 9728), and 30-day rotating refresh tokens with automatic reuse detection. Any client that speaks the spec works out of the box — including ones that haven't shipped yet.

FEATURE: Credit guarantee — charged or refunded, always
The mockup and scene generation pipelines have been unified behind a single orchestrator. Every credit either ends in a visible asset or is refunded — including failures from model errors, blob upload errors, and serverless invocation timeouts. A periodic sweeper backstops the rare case where the runtime is killed mid-pipeline, refunding within 5–10 minutes.

FIXES: Retired the legacy REST API
The `/api/v1/*` REST endpoints and the `api_keys` table have been removed. The MCP server now covers every read-only operation the v1 API exposed (plus the new mockup/scene creation), all over a single OAuth-authed endpoint instead of long-lived keys. If you had a script or webhook pointed at `/api/v1/*` with a `qk_live_*` token, switch it to the MCP endpoint.

Mockup Studio Overhaul, Asset Uploads, and Platform-Wide UI Polish

Mon, 13 Apr 2026 12:00:00 GMT

This cycle delivered a completely redesigned mockup studio mobile experience, a new asset upload system, and comprehensive UI standardization across the platform. Plus a new fullscreen image viewer, enhanced persona integration, and dozens of quality-of-life improvements that make the entire app feel more polished and responsive.

NEW: Direct asset uploads to your workspace
Upload image files directly to your workspace through a new drag-and-drop upload dialog. Supports PNG, JPEG, SVG, and WebP files up to 10MB each. Assets appear immediately in your Asset Gallery with automatic thumbnails and metadata. Access uploads from the Assets page header or the "Upload Assets" link on your workspace homepage.

NEW: Fullscreen image viewer with zoom and gestures
View any generated mockup, asset, or persona image in a dedicated fullscreen viewer. Features pinch-to-zoom, double-tap zoom toggle, drag-to-close gestures on mobile, plus mouse wheel zoom and keyboard shortcuts on desktop. Download images directly from the viewer or navigate back to their source context.

NEW: Complete mockup studio mobile experience
Rebuilt the entire mockup studio interface for mobile devices. The generate button stays sticky at the bottom, images center properly in portrait mode, prompt text displays as a clean 2-line preview that expands to a bottom sheet, and all touch interactions feel native and responsive. Camera positions, device types, and scene selection all work seamlessly on small screens.

FEATURE: Smart persona integration in mockup generation
Mockup generation now uses both profile photos and environment images as character references for better consistency. When you click "Mockup" from a persona page, the studio auto-applies that persona and displays context breadcrumbs showing which product and persona you're working with. Generated mockups include clickable product/persona links for easy navigation.

FEATURE: Standardized page layouts across the platform
Brands, Assets, and Mockups pages now follow a consistent layout pattern with sticky headers, unified search and filter controls, and matching content areas. Filter dropdowns use the same styled components, search bars have consistent placement, and coach mark info buttons provide contextual help where needed.

FEATURE: Enhanced persona chat reliability and feedback
Fixed the chat input becoming permanently disabled after multiple questions and added a visible typing indicator with animated dots when personas are generating responses. Chat sessions now handle unlimited length conversations with automatic message windowing to prevent token overflow while preserving full history in the UI.

FEATURE: Recent content carousels with actions
Homepage now displays 10 recent mockups and assets (up from 5) with hover actions for quick download and delete operations. Added "See All" cards when you have more items beyond the carousel display limit, making it easier to access your full library without navigating to dedicated pages.

FEATURE: Advanced mockup generation controls
Added device type selection (iPhone, MacBook Pro, iPad, Studio Display, and more), expanded camera positions including "over the shoulder" and "device close-up" angles, and enhanced lens options with macro and depth of field controls. Screenshot uploads now integrate directly into device selection for streamlined workflow.

FIXES: Mobile navigation and interaction polish
Fixed sidebar auto-collapse in the mockup studio, eliminated auth button loading flashes on marketing pages, improved mobile filter bar layouts that wrap properly on small screens, and resolved image loading issues across Assets and Mockups pages. Touch targets are larger and more responsive throughout the mobile experience.

FIXES: Image loading and performance improvements
Resolved broken image display on Assets and Mockups pages caused by Vercel Blob storage configuration conflicts. Added proper URL normalization for blob-stored images, improved masonry grid performance with CSS aspect ratios, and eliminated layout thrashing during image loading with skeleton states.

Asset Gallery, AI Logo Generation, Streaming Chat, and Mobile Overhaul

Mon, 30 Mar 2026 12:00:00 GMT

This cycle's headline: a new Asset Gallery that organizes every image in your workspace into a single browsable, filterable view. We also shipped complete AI logo generation, migrated Ask AI chat to real-time streaming, rebuilt the mobile experience, and made substantial performance improvements throughout.

NEW: Asset Gallery — every workspace image in one place
Browse all your persona avatars, environment images, AI-generated logos, and brand logos in a Pinterest-style masonry gallery. Filter by type (Images, Logos, Avatars), source (Persona, Brand, Upload), or product name. Click any image to open a full-screen lightbox with keyboard navigation, source links back to the original persona or brand, and one-click downloads. Assets are indexed automatically as you generate persona images and brand logos. [Learn about Assets →](/docs/assets)

NEW: Complete AI logo generation system
Generate professional logos from any brand profile with AI. The system creates 4 distinct logo concepts (1 credit), generates all concepts in parallel (2 credits), lets you select your favorite, then produces 4 variations: logo mark only, dark mode, colored background, and monochrome (1 credit). Each variation downloads as a high-quality PNG. Total cost: 4 credits for a complete logo design package. [Learn about Logo Generation →](/docs/brands/logo-generation)

NEW: Real-time streaming chat for Ask AI
Chat with personas and documents now streams responses token-by-token, eliminating timeout failures and delivering instant feedback. The new architecture uses the Vercel AI SDK for reliable streaming. Responses appear immediately as the AI writes them, suggested questions load reliably after each answer, and session recovery handles auth expiration gracefully. Performance improved from 20-30 second delays to 1-2 second first responses. [Learn about Persona Chat →](/docs/personas/chat)

NEW: Mobile experience rebuilt from the ground up
Completely overhauled mobile navigation with proper touch interactions. Mobile menus now have large tap targets, visual press states, and close automatically when you navigate. Added quick workspace navigation for logged-in users in the public mobile menu. Fixed hover states that don't work on touch devices, improved bookmark interactions, and eliminated phantom behaviors across the entire mobile experience.

FEATURE: Enhanced brand publishing workflow
Brand profiles now use the same sophisticated publish popover as persona sets and projects. Publish internally for workspace sharing, toggle community visibility, copy view-only links, and unpublish with granular control. The flow includes last-published timestamps and streamlined sharing options that match the rest of the platform. [Learn about Publishing →](/docs/community)

FEATURE: Major performance improvements across all pages
Homepage, community pages, and workspace sections now use Suspense streaming with skeleton loading states. Above-the-fold content renders immediately while data streams in asynchronously. Eliminated redundant database queries, moved workspace-only providers out of public routes, and added dynamic imports for heavy components in settings. PostHog analytics now loads lazily to reduce initial bundle size.

FEATURE: Brand audience section with persona linking
Brand profiles can now connect to persona sets through a dedicated Audience section. Link existing persona sets or create new ones directly from the brand page. The section displays target audience descriptions from brand analysis alongside persona avatar cards that navigate to individual persona profiles. Strengthens the connection between brand identity and customer understanding. [Learn about Brands →](/docs/brands)

FEATURE: Logo collection with multiple sources
Brand logo sections now aggregate logos from SimpleIcons (3000+ SVG brands), DuckDuckGo, Clearbit, Google Favicons, and extracted site assets. Each logo includes download buttons with format badges (SVG/PNG) and attempts direct downloads with CORS fallback. Added PDF brand package export with comprehensive multi-page layouts covering brand story, colors, typography, voice guidelines, and visual direction. [Learn about Brand exports →](/docs/brands)

FEATURE: Public brand pages with full content parity
Community brand pages now show complete brand profiles with all sections: messaging, story, colors, typography, voice, visual direction, logos, and brand audit. Free access to viewing content and downloading logos, with sign-up gates for exporting configs and forking brands. Enhanced SEO with JSON-LD organization schemas and richer meta descriptions. [Browse community brands →](/community/brands)

FIXES: Ask AI chat reliability and quality improvements
Fixed suggested questions not loading after streaming responses, eliminated input becoming permanently disabled, and resolved session expiration crashes. Restored full persona behavioral prompts after performance optimizations had degraded response quality. Added retry logic with timeouts for reliable infinite chat sessions and improved error handling throughout the chat flow.

FIXES: Mobile UX polish across platform features
Mobile sidebar navigation now provides immediate visual feedback on tap, closes properly when navigating, and hides hover-dependent bookmark buttons. Marketing mobile menu gained scrollable content, consistent press states, and workspace navigation links for authenticated users. Document editor AI tools dropdown now renders above other page elements with proper z-index layering.

Brand Identity System, API Launch, and Major UX Improvements

Mon, 23 Mar 2026 12:00:00 GMT

This two-week cycle brought our biggest feature yet: a complete brand identity system for generating and managing brand guidelines. We also launched our public REST API, redesigned the homepage into a proper marketing experience, and made substantial mobile and performance improvements.

NEW: Complete brand identity system
Generate comprehensive brand guidelines from any URL or description. Extract color palettes, typography, voice and tone, messaging frameworks, and visual direction principles. Each brand profile includes logos, brand story, audience targeting, and brand audit scoring. Share brand guidelines publicly or export as PDF, CSS variables, or Tailwind config. [Learn about Brands →](/docs/brands)

NEW: Public REST API for developers
Access all your Brand Stack data programmatically with our new REST API. Read personas, documents, projects, and task balances with API key authentication. Perfect for integrating with Slack bots, CI/CD pipelines, or custom tooling. Rate limited to 60 requests per minute with comprehensive documentation. [Read the API docs →](/docs/api)

NEW: Redesigned homepage as a marketing experience
The homepage is now a complete product showcase with concrete metrics (< 2 min generation, 6 personas per URL), interactive How It Works flow, and capabilities overview. Added social proof section, testimonial cards, and community showcase featuring real persona sets and brand profiles from the platform. [Visit the homepage →](/)

NEW: AI agent-friendly content negotiation
All public pages now support `Accept: text/markdown` for structured content consumption by AI agents and tools. Includes `/llms.txt` discovery file and automatic content formatting. Perfect for AI coding tools that need customer context or research automation workflows. [Learn about Agent Context →](/docs/personas/agent-context)

FEATURE: Mobile experience overhaul
Completely rebuilt mobile navigation with proper tap targets, press states, and automatic menu closing on navigation. Added workspace links to public mobile menu for logged-in users. Fixed sidebar interactions, bookmark buttons, and eliminated phantom hover states that don't work on touch devices.

FEATURE: Performance optimizations across the board
Homepage loads instantly with Suspense streaming — above-fold content ships immediately while community section streams in with skeletons. Workspace pages eliminated redundant database queries, moved workspace-only providers out of public routes, and added lazy-loading for PostHog analytics. Settings pages now use dynamic imports and proper loading states.

FEATURE: SEO improvements for all public pages
Added structured JSON-LD data, canonical URLs, and enhanced meta descriptions to personas, brands, projects, and documents. Updated sitemap generation and robots directives. Public pages now have rich snippets potential and better search visibility for customer research queries. [Explore the community →](/community)

FEATURE: Enhanced developer experience
New Developer settings page for API key management with admin controls, key creation/revocation, and usage monitoring. Added comprehensive API documentation with endpoint reference, authentication guide, and integration examples. [Get your API key →](/docs/api)

FIXES: Content negotiation and mobile polish
Fixed drag-and-scroll carousel interactions on desktop and mobile with proper pointer event handling. Improved settings page responsiveness, developer key display truncation, and eliminated layout overflow on small screens.

FIXES: Brand system reliability and security
Enhanced JSON parsing robustness for AI responses, implemented comprehensive input validation for colors and fonts, and added proper CSRF protection for all brand server actions. Improved error handling and user feedback throughout the brand creation flow.

Homepage Redesign, Mobile Polish, and Reliability Improvements

Mon, 16 Mar 2026 12:00:00 GMT

This week we redesigned the homepage into a full marketing experience, polished mobile UX across the app, and made persona generation significantly more reliable. We also improved page performance and added smart credit pack purchasing.

NEW: Homepage redesign with value communication
The homepage is now a complete marketing page that clearly explains what Brand Stack does and why it matters. New sections include How It Works (3-step flow), Capabilities (4 key features), and Built for Your Team (specific use cases for Product, Marketing, and Sales). The hero leads with "Know your customers before you build" instead of just describing features. [Visit the homepage →](/)

NEW: Redesigned credit pack purchase experience
Credit pack cards now lead with task count as the hero number, show dollar savings explicitly with strikethrough pricing, and highlight Popular and Best Value options. Makes it easier to understand value and choose the right pack size. [Learn about credit packs →](/docs/billing)

FEATURE: Mobile experience improvements
Mobile menus now close automatically when you navigate to a new page. The Settings link opens a proper settings menu instead of jumping straight to Preferences. Input fields no longer trigger browser zoom on focus, and share popovers center properly on mobile screens.

FEATURE: More reliable persona generation
Persona recommendations now succeed on the first attempt far more often. We increased model timeouts, added smarter retry logic, and softened quality assertions that were rejecting valid output. The new system also survives tab switching — if you come back to a failed generation, it automatically retries. [Learn about creating personas →](/docs/personas/creating-personas)

FEATURE: Faster page loads throughout the app
Projects, Personas, and Documents pages now skip unnecessary database queries when the credits system is active. This eliminates up to 12 redundant queries per page load, making navigation noticeably snappier across the workspace.

FEATURE: Enhanced SEO for all public pages
All public pages now have proper title tags, meta descriptions, Open Graph tags, and canonical URLs. The sitemap includes pricing and changelog pages. Better search visibility for users discovering Brand Stack. [Explore the community →](/community)

FIXES: Members page updated for workspace billing
The members page no longer shows per-seat pricing or paid/free badges since all plans include unlimited users. Added an inline invite form and simplified role descriptions for the credits era.

FIXES: Credit pack modal and subscription updates
All upgrade buttons work independently (no more disabling all buttons when clicking one). Free plans show "Free" instead of "$0/mo". Subscription management reflects the new task-based language throughout.

Faster Pages, Smarter Mobile, and a More Reliable Persona Engine

Sat, 14 Mar 2026 12:00:00 GMT

This week we focused on speed and polish. Pages load faster with fewer unnecessary queries, mobile navigation works the way you'd expect, and persona recommendations are significantly more reliable. We also redesigned the credit pack purchase experience and shipped a wave of mobile improvements.

NEW: Redesigned credit pack purchase cards
The "Buy More Tasks" modal now leads with the task count as the hero number, shows dollar savings explicitly, and highlights the best-value options. Discounted packs display strikethrough pricing so you can see exactly what you're saving. [Learn about credit packs →](/docs/billing)

NEW: Mobile settings menu
Tapping Settings on mobile now opens a full settings menu instead of dropping you straight into Preferences. Pick exactly which settings page you want before navigating.

FEATURE: Faster page loads across the app
Projects, Personas, and Documents pages now skip unnecessary database queries when the credits system is active. This eliminates up to 12 redundant queries per page load, making navigation noticeably snappier.

FEATURE: Smarter document table performance
The documents table now uses an indexed lookup for project associations instead of scanning all records for every row. Larger workspaces will feel this improvement the most. [Learn about Documents →](/docs/documents)

FEATURE: More reliable persona recommendations
Persona generation now succeeds on the first attempt far more often. We increased model timeouts, added smarter retry logic, and stopped rejecting valid personas over minor vocabulary differences. The new loading experience shows six AI agents working together with live progress updates. [Learn about creating personas →](/docs/personas/creating-personas)

FEATURE: Persona recommendations survive tab switching
If you switch tabs while personas are generating, the request keeps running in the background. Come back and your results are waiting. No more lost progress from checking another tab.

FIXES: Mobile menus close after navigation
All mobile sidebar menus now close automatically when you tap a navigation item. Previously, the page would change behind the menu but the menu stayed open.

FIXES: No more zoom on mobile inputs
Tapping input fields on mobile no longer triggers browser zoom. The keyboard appears cleanly without shifting or scaling the page.

FIXES: Share popover works on mobile
The share and publish popover on persona pages now centers properly on mobile screens and no longer shows stuck tooltips on touch devices.

FIXES: Members page updated for workspace billing
The members page no longer shows per-seat pricing, paid/free badges, or seat counts. All plans include unlimited users, and the page reflects that with a cleaner inline invite form. [See billing details →](/docs/billing)

Task-Based Credits, Smarter Billing, and a Redesigned Subscription Experience

Mon, 09 Mar 2026 12:00:00 GMT

We replaced seat-based pricing with a task-based credit system that gives every workspace unlimited users. Choose a plan, use tasks for AI actions, and top up anytime. The new Subscription page puts everything in one place.

NEW: Task-based credit system
Every AI action in Brand Stack now costs a clear number of tasks. Generate a persona (3 tasks), create a document (1 task), chat with a persona (0.1 tasks), or use AI editor tools (0.5 tasks). No more guessing about token limits or seat counts. [See the full pricing breakdown →](/docs/billing)

NEW: Five flexible pricing tiers
Choose the plan that fits your team. Free (10 tasks to explore), Starter ($9/mo, 30 tasks), Pro ($19/mo, 70 tasks), Agency ($49/mo, 300 tasks), or Business ($149/mo, 1,000 tasks). Every plan includes unlimited users. Save 20% with annual billing. [Compare all plans →](/docs/billing)

NEW: Credit pack top-ups
Need more tasks mid-month? Buy credit packs starting at $5. Bulk packs offer up to 37% savings. Purchased credits never expire. [Learn about credit packs →](/docs/billing)

FEATURE: Redesigned Subscription page
Upgrade, downgrade, cancel, or resubscribe directly from the new Subscription page. See your remaining tasks, plan details, next billing date, and manage everything in one place. No more jumping between separate billing and plans pages.

FEATURE: Instant subscription management
Cancel and resubscribe with one click. Downgrade through the Manage Subscription modal with a clear summary of what changes. Your subscription credits stay active for 30 days after cancellation so nothing is lost abruptly.

FIXES: Real-time credit updates
Your task balance now updates instantly after purchases, upgrades, and subscription changes. No more refreshing the page to see updated numbers.

Persona Skills Export and Community Publishing

Mon, 23 Feb 2026 12:00:00 GMT

Export your personas as structured skill files for coding agents, and publish your work to the Brand Stack community for others to discover and learn from.

NEW: Persona Skills export
Export your persona set as a structured directory of skill files, ready to drop into your coding agent's workspace. Each persona becomes a detailed reference file that AI coding assistants can use to build with real customer context. [Learn more about Persona Skills →](/docs/personas/skills-export)

FEATURE: Community publishing
Share your personas, projects, and documents with the Brand Stack community. Published work is discoverable by other users and helps build a shared library of customer understanding across teams and industries. [Learn more about Publishing →](/docs/community)

FEATURE: Improved persona chat
Chat with any persona to pressure-test ideas, validate messaging, or explore customer perspectives. Web search toggle lets personas ground their responses in real-time information when you need current context. [Learn more about Persona Chat →](/docs/personas/chat)

FIXES: Smoother onboarding flow
The new workspace setup experience is faster and clearer, with better guidance through the URL-to-persona pipeline. First-time users can generate their initial persona set in under two minutes. [Get started →](/docs/getting-started)