Glossary
Every term in the new MindBacklog, in plain words
One to three sentences each, grouped by place. Old names are listed too, so a search for one still lands here.
The words that matter most
- Mind
- The AI in MindBacklog. Mind drafts, reads and answers; you decide. It shows its sources, labels what is yours and what is public, and says when it has nothing rather than guess. More
- Ask Mind
- The place to ask Mind about your product: a page with your chats and a panel on every page scoped to what you are looking at. Answers mark which sources are yours and which are public. When your data has nothing on a question, Mind says so, and what would change the answer. More
- Feedback
- One piece of what a user told you, in their words, from the widget, a form, the portal, email, Slack, an app store review, the API or something you dropped in. Learn › Feedback lists every piece with Mind's one-line read under it. More
- Problem
- A problem worth solving, made from feedback and ranked by how many of your own users have it. Each has a key (P-1, P-2), the quotes behind it and your call on it. Your users rank first; public posts and competitor reviews are shown, never counted. More
- Belief
- Something that must be true for your product to work, tracked on Learn › Questions. Each belief sits on the evidence ladder and climbs only when someone did something. The word replaces "assumption". More
- Evidence ladder
- Six rungs a belief climbs, weakest to strongest: Praise, Past pain, Asks for access, Gives time, Makes an intro, Pays. A belief moves right only when someone did something; praise never counts as proof, and your own words never count. More
- Build Pack
- One Markdown document per feature that your coding agent pulls or you copy. It carries why it matters, who it's for, the metric, and the must-haves with acceptance criteria. It also carries what is out of scope, known issues, the evidence as counts and short quotes, and what already exists. Every section names its source, and the only numbers are counts. More
- Product map
- What your product does by area and how it is built, written by your coding agent over MCP and approved by you. Each capability shows Live, Partly live, In progress, Scoped or Idea. Your coding agent writes the product map. MindBacklog never reads your code. More
- Initiative
- The level above features: the why, a time budget and a prediction. Keys I-1, I-2. A feature can sit under several initiatives, and one timeline shows every initiative with its features. The word replaces "objective". More
- Product intelligence
- Keeping what your users said, what you decided and why, and what you shipped in one record, so every product decision can show its evidence and be checked afterwards. The practice MindBacklog is built around. More
- People
- Everyone who paid you, wrote in or talked with you, one page each: what they pay, what they said, every touch with its date and source, and whether you may contact them. Keys C-1, C-2. A research base, not a CRM: no enrichment, no outside lookups, no contact lists. More
Learn
- Drop anything
- The one way to hand Mind something by hand (⌘D): a thread, a transcript, a CSV, a link. Mind works out what it is and shows where it will go before filing it. Replaces Quick Feedback, Log Signal and Quick Dump. More
- New patterns
- A view on Problems: what Mind spotted in feedback filed under no problem yet. Each pattern waits for you to make it a problem, keep it as a tag or leave it. Replaces "trending themes". More
- Check a request
- Type an ask ("we need a mobile app") and test it against what your users actually said, for and against. The verdict (Strong, Some, Weak, No evidence either way, or Not enough feedback) is counted from people, never written by the model. More
- Question
- An entry on Learn › Questions: either a belief to test, or something you would look up before a decision. The tab shows what you don't know yet, with the riskiest belief flagged. More
- Assumption
- Now called a belief. The word was retired in the app because an assumption sounds like something to accept, and a belief is something to test. More
- Praise (rung 1)
- They said something nice: "I'd use that", "I'd pay for that". It costs nothing and never moves a belief. An answer to your own leading question is praise, whatever it sounds like.
- Past pain (rung 2)
- They described a specific time the problem hurt, unprompted. The first rung that counts, because they did something, even if the something was suffering.
- Asks for access (rung 3)
- They asked to get in before it exists. As far as a conversation alone can carry a belief.
- Gives time (rung 4)
- They showed up: a second call, a trial they used, a test they took part in. Logged by you, with the date.
- Makes an intro (rung 5)
- They introduced you to someone and put their name beside yours. Logged by you, with the date.
- Pays (rung 6)
- Money changed hands, before or after the product exists. Reached only through a test whose pass line was set before it started.
- Evidence step
- One rung of the evidence ladder. A belief's evidence step is the strongest rung it has reached, and the weekly review shows the quote or action that moved it. More
- Conversation
- A call or meeting you captured from a transcript, with consent recorded. Keys V-1, V-2. Mind reads only the other person's turns, pulls their exact words, leaves out yours and marks what you prompted. MindBacklog does not record calls. More
- Call guide
- Your interview questions, kept in sections, that you pick when you log a call. The questions are checked against six rules for leading questions before the call. More
- Test
- An experiment on a belief whose pass line is set before it starts and cannot be edited afterwards. Only the deadline can move, with a logged reason. A pass is the only way a belief reaches Pays. More
- Pass line
- The number a test must reach by its deadline, written and locked before the test starts: "10 pre-payments by 15 Oct". Results that come in never move a belief by themselves; the verdict against the line does. More
- Decision
- A logged call: build, test first, not now, no, or a pick between options. It keeps its reason, its topic, whether it is a one-way door, who decided and a review date, with the evidence frozen as it stood. Learn › Decisions is the log. More
- Who it's for
- Your customer segments as living profiles: the main segment, candidates and groups outside your customers. Every line says how it is known: proven from people and payments, or still a belief. "ICP" is fine in prose; the screen is called Who it's for. More
- Coaching
- Small checks on your research habits, mostly by rule and free. Your interview questions before a call, and how the call went after it: who talked more, when you first pitched. Blind spots in your beliefs, and price guardrails from what customers actually said about money. More
Build
- Feature page
- One page from why to what to build. The state (Idea, Shaping, Ready to build, In build, Shipped, Parked), and why it exists with who asked in their words. Then Mind's read, Ready to build?, the docs and the Build Pack. Replaces the old item page with its eight tabs. More
- Acceptance criteria
- The checks that define done for a story, written so someone can tick them: "Given an unpaid invoice, the client can pay by card." They travel with the story into the Build Pack. More
- Ready
- A story's state once it has acceptance criteria and its feature's spec is approved. A story is Ready only when it has them; the board will not move one into build without them. More
- Story
- A unit of work under a feature, with its acceptance criteria, status and owner. Build › Stories lists every story and bug. Replaces the Tracker. More
- Spec interview
- Seven plain questions from which Mind writes the PRD and the stories. Who uses it first, what happens today, what must happen after, what it must never do, what data it touches, the smallest slice, how you'll know it works. Your answers are the authority; the target is always yours. More
- PRD
- The product requirements document for one feature. It is one document among a feature's docs, drafted from evidence and set Draft, In review or Final. Final on the PRD approves the spec. The old "Living PRD" is now simply the PRD. More
- Documents
- Every PRD, brief, help article and launch plan across your product in one list, each saying where it came from and whether feedback has arrived since it was last edited. Replaces AI Artifacts. More
- Stale
- A document is stale when feedback on its feature arrived since the document was last edited. The row says how many pieces. It is a prompt to read them, never a verdict that the document is wrong. More
- Time budget
- How long an initiative is worth before you stop: "Day 4 of 21". Set by the owner; health always shows with a reason. The word replaces "appetite". More
- Prediction
- "I expect {metric} to reach [your number] by {date}", written by the initiative's owner before the build. Mind fills in the metric, today's value and the check date. The number you expect is always yours. More
- Results
- Build › Results: did it work? On the check date, one outcome review per prediction: the prediction in the owner's words, the result from its source with n, before and after, and Mind's read, which never says whether it worked. You mark Worked, Didn't, Mixed or Too early. More
- Outcome review
- The review Results schedules for a prediction's check date. An initiative without a written prediction gets no review. More
- Publish check
- Build › Publish check: what changed since you last published, from your GitHub commit titles and merged pull requests. Also stories and bugs that moved, documents set Final, what each changed feature makes public, and GitHub's secret-scanning alerts. It lists; you confirm. MindBacklog never reads your code. More
- Plan
- Build › Plan: about ten plain questions, then the foundations a senior engineer would insist on before launch, each with its reason and source, plus the questions to send your developer. The plan from rules is free; tailoring it to your own features uses AI credits. More
- Launch
- Build › Launch: one launch (L-1, L-2) with its shipped features, its day and its channels. A readiness check that never blocks, the launch kit, the people to tell, reminders and lessons. MindBacklog never sends or posts. More
- Launch kit
- Four posts Mind drafts for a launch (Product Hunt, Hacker News, X and LinkedIn) from what your product says about itself and the done stories. A line about something not built yet is flagged; a post with a number or quote not in your data is dropped. You post them. More
- Tell the people who asked
- When a feature ships, the list of everyone who asked for it, with their own words, and who can't be told and why. Mind drafts a note to each; you send it. In the app the card reads "Tell the N who asked". More
- Evidence grade
- On every feature: Validated problem, Assumption only or No evidence, counted from your feedback, beliefs and tests, with the reason one click away. No AI writes it. A grade does not prove a solution will work. More
Today
- Today
- The home page: one next move every day, with its why, chosen by rules from your data. Plus Mind's one question, what changed since your last visit, moves this week and people to talk to next. Replaces the Dashboard. More
- Journey
- Ten stages from problem to product-market fit (Problem, Customer, Solution, Positioning, Demand validation, Build v1, Launch, Learn, Retention, PMF), each with a gate. Mind never moves a stage or blocks your work. Past PMF the tab becomes Health. More
- Stage
- One of the ten Journey stages. You pass a stage when its gate is met, or skip it with a reason that is logged as a decision. More
- Gate
- The checks that say a stage is done. Each is either checked by data (Mind counts it from what is stored, with its source) or you confirm (a judgement only you can make, confirmed with a note). More
- Check-in
- A two-minute monthly check on Journey: who decides what gets built, where the product is, the biggest risk, and what the data doesn't show. Free, by rule, with What Mind sees beside each question. More
- Weekly review
- Friday: what moved this week, your calls checked against the lines you set, decisions due, the week's number against its target, and an update Mind drafts and you send. Nothing is sent from MindBacklog. More
- Digest
- What moved, by email and in the bell, daily or weekly, with one paragraph of Mind's read built from the week's events. Replaces the daily brief. More
Market
- Positioning
- Market › Positioning: what you have confirmed about your product, Mind's drafts to check, the Lean Canvas drawn from your own data, and the price guardrails. Each line is versioned. More
- Competitor watch
- Watch a competitor and Mind reads everything it can find about them: site, pricing, help center, changelog, status page and app reviews. Public pages only; it respects robots.txt. Every read is dated and sourced, and if something isn't in their docs, Mind says so rather than claim they lack it. More
- Industry
- Market › Industry: your industry profile, and every Monday what moved in your market that touches your buyer, your problem or the rules, each item dated and sourced. None of it counts as your evidence. More
- Sources
- Market › Sources: every input in one list (widget, web forms, portal, inbound email, REST API, Slack, app store reviews, pastes) with what each found, kept and ignored. Reddit is off by default for new products: paste a Reddit or Facebook thread instead. More
- Outside research
- Mind's search of public pages for a question of yours. Outside numbers come as ranges with their sources, marked as estimates, and never move a belief. It counts toward the monthly AI allowance. More
Grow
- Grow
- The place that appears past product-market fit, or from ten paying accounts: Revenue, Retention and Growth, read-only from your own Stripe and PostHog. Under the threshold it says "Not enough data" instead of a verdict. More
- The 40% test
- The product-market fit survey: "How would you feel if you could no longer use the product?" The share who answer very disappointed, leaving out those who no longer use it. Shown with n and a range, by segment. Forty percent or more holds, from thirty answers. More
- Health check
- One of four readings per main segment: Still wanted (the 40% test), Segments stay (60% still paying in month three), Expansion beats churn (net revenue retention of 100% or more), Repeatable channel. Each shows its value, range, n and source. More
- Focus
- The one health check, for one main segment, that matters most now. It moves only with a reason, and every move is a decision in the log. More
- Slipping
- A health check that held and then fails for two weeks of readings. It goes to the top of Today and into the digest. More
- Revenue at stake
- On Grow › Revenue: your open problems and initiatives ranked by the MRR of the paying accounts whose feedback is linked to them. Counts from your own Stripe, never a forecast. More
Settings, plans and AI use
- Connections
- Settings › Connections: every tool you connect lives in one place, with two switches each. Sync brings what happens into MindBacklog; Ask lets Mind look a tool up live while answering, never stored and never counted as evidence. More
- Your AI tools
- Settings › Your AI tools: the coding agents and AI tools connected to MindBacklog over MCP, with the steps to connect each and a first prompt to try. On Growth and Scale. Replaces "Developer settings" and "MCP tokens". More
- MCP
- The Model Context Protocol, an open standard that lets an AI tool read another app. MindBacklog's MCP server lets Claude Code, Cursor and other MCP clients read your features, Build Packs, stories, documents, problems, feedback and map, and write only as far as you could. Nothing can be deleted over MCP. The tools make no AI calls and cost no credits. More
- AI credits
- Credits that pay for the AI work you start yourself: Ask Mind, documents, Check a request, tailoring a plan, a battle card. Starter: 1,500 a seat a month, pooled. Growth and Scale: included. Credits come back when a job fails, is cancelled or finds nothing to use. More
- Signal credits
- One per piece of feedback that comes in: a CSV row, a form answer, a support ticket filed as feedback. 1,500, 3,000 or 7,500 a seat a month on Starter, Growth and Scale. More
- Monthly AI allowance
- The cap on Mind's background AI work each month: reading and filing new feedback, spotting patterns, each problem's read, the weekly watches. At the allowance, background work pauses until the first of next month and nothing is lost; what you start still runs. There are no weekly caps. More
- Workspace
- Your account, holding one or more products and your team. Replaces "Organization". Admins and editors take a seat; viewers are free. More
Names that changed
The new MindBacklog uses one set of names. If you knew the product by an older name, this is where it went. Old links redirect.
| Was called | Now called |
|---|---|
| Intelligence Hub | Problems |
| Insight | Problem |
| Raw Feed | Feedback |
| Trending themes | New patterns |
| MIND Engine | Mind |
| Signal Pipeline | Feedback (the sources and the list) |
| Signal | A piece of feedback |
| Concept Mode | Journey (the idea path and its stages) |
| Daily Brief | Digest |
| Dashboard | Today |
| Tracker | Stories |
| AI Artifacts | Documents |
| Living PRD | PRD (a document) |
| Objectives (OKRs) | Initiatives |
| Competitive Intelligence | Competitors |
| Quick Feedback, Log Signal, Quick Dump | Drop anything |
| Integrations, Developer settings, MCP tokens | Settings › Connections and Your AI tools |
| Organization | Workspace |
| Product Intelligence Platform | MindBacklog, the brain for your product |
One set of words. One record.
Start from an idea, your first users or a live product, and see each of these on your own data.
- 14 days free
- No card needed
- Viewers are free
