Every ship,
transparently.

Confluence connector — Ask Mind reads your wiki live.

MindBacklog's MCP server has always pointed outward — your AI clients querying your product intelligence. It now runs both directions. Connect your team's Confluence wiki and Ask Mind searches it live — specs, design docs, runbooks, decisions — citing the real pages alongside your customer signals and roadmap. Nothing is imported or indexed; every search runs against Confluence the moment you ask, through Atlassian's official remote MCP server, with a scoped API token you control and revoke anytime. Read-only, Admin-only to connect, available on Growth and Scale plans. Confluence is the first connector — Amplitude is next.

Also in V2.1: Integrations moved out of Product Settings into its own page in the sidebar, split into two tabs — Connectors (MCP-based, reading your own tools live) and Native Apps (Jira Cloud's bidirectional sync, Slack feedback capture). Old links redirect automatically.

Intelligence Hub, Competitive Intelligence, and a simpler two-stage flow.

The biggest release since v1.0. Radar is now the Intelligence Hub — a morning briefing with a daily AI brief, Signal Trending (see which themes are getting louder week-over-week), AI cluster suggestions that turn into Roadmap features with one click, and a refined raw signal feed. Competitive Intelligence is a dedicated page: sentiment cards per competitor, gap & opportunity matrix, battle cards, switching signals, and a live Market Moves feed. Ask Mind is now a persistent side panel accessible from any page via the top bar.

Also in V2: Quick Feedback capture (paste any messy notes from the top bar — AI splits and classifies them), Daily Intelligence Brief by email, RICE as the default scoring framework, and the Ideas bucket removed in favour of a clean two-stage Intelligence Hub → Roadmap flow.

Ask Mind 2.0 — web research & smarter PM coaching.

Ask Mind can now search the public web for competitor intel, industry benchmarks, and PM frameworks — every external claim comes with inline citations tied to real source URLs. It applies the right PM framework automatically (PRD structure, RICE prioritization, coaching mode), remembers what you discussed earlier in a conversation, and always knows your live product state — feature counts, recent signal volume, ship velocity, top theme. Honest by default: if evidence is thin, it says so.

Native Zapier integration.

MindBacklog is now a native Zapier app — pick it from Zapier's app picker and connect any of 7,000+ supported tools to your feedback pipeline without configuring a custom HTTP webhook. Every Zapier-sourced signal runs through the same 4-phase AI pipeline as native channels. Currently in private beta during Zapier's app review — email [email protected] to be added.

Reddit monitoring — live.

Track competitor subreddits, your product's own subreddit, or broader market topics. Reddit threads flow into the same Listen pipeline as App Store reviews — sentiment-tagged, lens-classified, and matched against your roadmap. When Competitor Discovery finds a verified subreddit for a competitor, it auto-creates the Reddit channel.

Stripe Managed Payments — global tax & more payment methods.

Billing moved to Stripe Managed Payments — Stripe acts as Merchant of Record handling VAT, GST, and sales tax across 80+ countries automatically. Card-less 14-day trials, Apple Pay / Google Pay / Stripe Link, prorated upgrade transparency with exact charges shown, promo codes at checkout, and a self-serve billing portal in Settings → Billing.

MCP server — bring your own AI client.

Connect Claude Desktop, Claude Code, Cursor, ChatGPT, or any MCP-compatible AI client directly to your MindBacklog workspace. 18 read-only tools across signals, features, roadmap, objectives, teams, and the product knowledge base. Bearer-token auth, scoped per org or product, up to 10 active tokens. Drop-in JSON config snippets for every major client in Settings → Developer. Available on Growth and Scale plans.

Competitor Discovery Engine.

The MIND Discovery Engine scans the web and app stores to find, rank, and track your competitors using your product context — including ones you haven't heard of. Approve a suggestion to instantly create App Store, Google Play, and Web Scraper tracking channels with one click.

Listen · Decide · Build — the full platform is live.

The complete Listen → Decide → Build framework ships as v1.0. Universal ingestion pipeline with native channels, Slack, app reviews, web scraping, REST API (POST /api/channels/v1/ingest), and Zapier/Make/n8n support. Evidence-based WSJF and RICE scoring. Ask Mind co-pilot with dozens of specialized PM tools — query, create, prioritize, generate PRDs and user stories, schedule, score, and more via natural language.

Also in v1.0: Concept Mode for pre-launch intelligence, embeddable Feedback Widget, Public Portal for customer feature requests, Tracker (kanban for user stories and bugs), AI content generation for PRDs and release notes, live bidirectional Jira sync, and in-app notifications. Stripe billing with 14-day free trial.

MIND Context Engine v2.

Multi-level summary strategy — master summaries, source summaries, and chunk embeddings working together for richer context with fewer tokens (≈40% lower AI cost per query). Source relevance validation filters out competitor mentions and generic industry content before they reach context. Summaries now update incrementally as new sources are added. pgvector replaces the external vector DB; Redis pipeline handles source ingestion 3× faster.

Feature scoring — WSJF and RICE engines.

Every feature gets automatic WSJF (Cost of Delay, User/Business Value, Time Criticality, Risk Reduction) and RICE (Reach, Impact, Confidence, Effort) scores, each grounded in your linked customer signals. Revenue impact estimation per feature. Scores update as MIND learns more. Unified side panel with tabbed navigation and horizontal score visualization bars.

AI signal clustering.

Raw signals from all sources are automatically clustered into distinct, actionable features in the Radar — "search is slow," "I hate the search," and "need better filters" become one feature with links to every original source. Cluster confidence scoring, manual merge & split (MIND learns from corrections), improved Radar feed sorting by source, date, sentiment, and processing status.

Multi-source ingestion pipeline.

Apple App Store reviews, Google Play ratings, inbound email (unique @mindbacklog.net addresses with real-time IMAP processing), and customer web forms — all flowing into a single unified Radar feed with source badges, timestamps, and sentiment indicators. Background processing via Laravel + Redis queue handles rate limits and retries gracefully.

Platform launch.

Email + password and Google OAuth authentication with multi-organization RBAC. Product workspace creation — upload docs, provide a URL, answer clarifying questions, MIND builds initial context. Three-bucket architecture: Radar → Ideas → Roadmap. Core stack: Laravel backend, PostgreSQL with pgvector, Python AI services, Redis queues, Docker containerization.