Changelog

New

• Image variables — variables can now hold an image, not just text. Upload your logo, headshot, or product photo once in Settings → Variables, and reuse it on every result and landing page with a single {{ custom:key }} token. It drops straight into an image tag — <img src="{{ custom:brand_logo }}"> — and swapping the image in Settings updates every page that references it, with no rebuild. Read the guide →
• Variable palette on the canvas — open a Page node and its settings panel now lists every variable you can use — custom values, brand images, and the product's form fields — each as a click-to-copy token. No more remembering keys.
• Claude sees your variables — ask the connector "list my custom variables" at any time and it returns every token and image, ready to place. Say "use my logo in the header" and Claude wires it into the page for you (get_variables now returns your custom values and images).

Improved

• Redesigned Variables settings page — a clearer, beginner-friendly walkthrough of what variables are, the difference between text and image variables, and the three ways to use them (copy a token, pick one on the canvas, or ask Claude).

Fixed

• Workspace image upload — uploading an image variable from Settings → Variables no longer fails with a "Product ID is required" error; workspace-level assets now save correctly (and .svg logos are accepted).

New

• Design in Claude Design — bring pages designed in Claude's visual canvas into Productised. Keep the connector in the same chat and just say "use this as my output page" — Claude produces the right format, wires your product's tokens, and saves it. The connector now reads your product's blanks up front (get_variables + get_design_system) and offers to design with you or take a Claude Design frame. Read the guide →
• Productised vs Claude Artifacts — a new explainer on why a hosted, data-capturing Productised product is the right tool for lead-gen funnels, where a one-off Artifact isn't. Compare →

  Improved

  • MCP connection test — connecting an external server (HubSpot, Notion, Linear, Slack, etc.) now performs a full handshake, so valid credentials are verified correctly against modern stateful servers instead of failing the check.
  • Documentation search in the connector now reaches the live docs reliably.

  Fixed

  • Claude Design offline export guidance — the "Export → Standalone HTML (offline)" file is a React bundle, not static HTML; importing it produced a blank page. The connector now flattens designs to static HTML automatically and warns against the raw offline file.
  • Plan generation no longer hangs when the AI service is busy — it returns a clear "try again shortly" message.

New

• Brain Sources tab — upload PDFs, TXT, MD, and CSV files or paste URLs directly into Brain. Files are indexed into your workspace knowledge base and drawn on in every Brain conversation.
• Prompt chaining on the canvas — chain multiple AI Product nodes in sequence. Pass outputs from one node to the next using {{ node:varName }} syntax in any downstream prompt or outcome.
• Scoring Node enhancements — define named persona profiles with custom score thresholds. Each persona can route to different outcome content, a different CTA, and an optional redirect URL.

Improved

• Variable picker updated to include chained node outputs alongside form fields and custom variables.
• Brain Chat response quality improved when Sources are active — citations now surface inline.

Fixed

• Canvas edge reconnection occasionally failed to persist after save. Resolved.
• Scoring Node threshold validation allowed overlapping ranges. Now enforced on save.

New

• Brain — workspace-level AI assistant to help you decide what to build and how to position it. Accessible from the main nav; not tied to any individual product.
• Brain Memory — automatically saves context from every conversation so you never have to re-explain your business, audience, or methodology.
• Brain Ideas — BUILD_THIS cards surface build-ready concepts from your conversations. Open any card directly on the canvas to start building immediately.
• Conversation tier limits — Free (1 Brain conversation/month), Starter (5/month), Pro (unlimited).

Improved

• Workspace onboarding updated to introduce Brain in the first session.

New

• Funnel view in the gallery — organise and filter your products by funnel tier: ToFU, MoFU, or BoFU. Assign a tier to any product from its settings.
• Product colour labels — tag products with a colour for visual organisation at a glance.
• Compact and Visual view modes — switch between a dense list view and the card grid depending on how many products you're managing.

Improved

• Gallery filter bar redesigned — filter by funnel tier, label, and status simultaneously.
• Product card now shows funnel tier badge alongside status.

New

• Scoring Node — add a Scoring Node to the canvas to evaluate collected responses and produce a numeric score. Three modes: AI Judgement, AI Scorecard, and Advanced Manual.
• Per-field score weights — in Advanced Manual mode, assign a weight to each collected field so the scoring reflects what matters most to you.
• Persona profiles — define named segments (e.g. "Ready to Scale", "Building Foundations") with custom content, CTAs, and optional redirects. Each persona activates when a response falls within its score threshold.

Improved

• AI Scorecard mode now accepts a rubric description alongside the numeric range, giving the AI clearer grading criteria.

Fixed

• Outcome pages occasionally showed raw variable syntax when a field was left blank. Now falls back to an empty string gracefully.

New

• Brand Voice (formerly Offer Profile) — captures your positioning, methodology, credentials, target audience, and guardrails. Feeds every AI Product in your workspace.
• AI Brand Voice builder — a short guided conversation that helps you build and refine your Brand Voice without starting from a blank form.

Improved

• Brand Voice now feeds Brain conversations as well as product outputs — consistent context everywhere.
• Renamed throughout the app from "Offer Profile" to "Brand Voice" for clarity.

Fixed

• Brand Voice changes not reflecting in active products without a manual save. Resolved.

New

• Connections Power — send lead data to external tools automatically after a product is completed. Supported destinations: Gmail, Google Sheets, HubSpot, Slack, Airtable.
• Zapier integration — trigger any Zap on product completion using the structured webhook payload.

Improved

• Webhook payload restructured — now includes the full response object: all collected fields, AI outputs keyed by node variable, timestamp, and product metadata.
• Connections setup moved to a dedicated canvas node so the data flow is visible and editable alongside your other nodes.

New

• Rich page builder for Report and Document outcomes — drag sections, add blocks, and arrange your outcome layout visually.
• Variable browser — search and insert any collected field, AI output, or custom variable into your outcome content using {{ category:key }} syntax. Preview live values from the latest response.
• Theme presets — choose from a set of pre-built visual themes for outcome pages, then customise from there.

Improved

• Report and Document outcomes now inherit Brand Voice styling defaults on creation.
• Outcome preview loads faster — no longer requires a full page reload when switching between responses.

New

• Research Power — give your AI Product live web search capabilities. Useful for products that need current data, market context, or competitor information.
• Advisor mode — after the collection phase ends, the AI shifts into an open-ended Q&A conversation. Respondents can ask follow-up questions and get personalised answers based on everything collected.
• Knowledge Power URL indexing — alongside file uploads, you can now paste a URL and have it indexed into your product's knowledge base.

Improved

• Knowledge Power file processing now handles larger PDFs more reliably.

New

• Productised public launch — the canvas, AI Product node, Knowledge Power, Report and Document outcomes.
• Collection goals — 10 field types available: text, email, choice (single and multi), scale, yes/no, image selection, and more.
• Decision model — two modes for how the AI evaluates responses: AI Judgement (holistic) and AI Scorecard (rubric-based).
• Product deployment — share via link or embed code. Product-level settings for visibility and email collection.