Reference

Glossary

The terms Productmind uses, in alphabetical order. Each one links to the page that explains it.

TermMeaning
Access tokenThe token a script or tool sends with every API request. It lasts an hour and acts as the person who authorized it.
AccountA company your feedback comes from. Contacts belong to accounts.
Action pointA follow-up from an internal meeting's summary, which you can assign to a participant and tick off.
Activity panelThe panel on the right of every page, with the Support center, Notifications, Integrations and PM agent tabs.
AdminA built-in role that manages the organization, its members and settings, with full access to all product work.
AgentOne of Productmind's AI agents, which chat, write specifications, run interviews and more. You customize them on the Agents page.
Agent contextStanding background about your product that the PM agent, the interview agent and the generators read. Coding sessions and scans don't read it.
Agent definitionAn agent of your own, or a customized standard one, with its instructions, model and tools, for an automation to run.
Agent offlineA session status: the machine running the session's agent went offline in the middle of a run. The session continues by itself when the agent is back, or you cancel it.
Agentic interviewsThe Feedback page where an AI interviewer talks to your customers by email or voice, and to visitors in the feedback widget. Its work is organized in researches.
Agentic readinessA scan that rates how well a repository is set up for AI coding agents.
AI assistant (MCP)An AI tool of your own, such as Claude Code, connected to Productmind over MCP.
AI creditsThe unit AI usage is measured in. Every model's tokens use AI credits in proportion to what they cost, a coding session in Productmind's cloud also uses 500 AI credits per web search, and Productmind's agents and coding sessions draw on the same AI credits, at one price. Coding sessions on your own machine use none of them.
AllowanceWhat your plan includes of something each billing period, such as its included AI credits. What you don't use doesn't roll over. Past it, overage decides whether work continues.
ApplicationA code repository registered with Productmind, with its services and how to run them. Its ID starts with APP.
Architecture mapA map of how your applications depend on each other, drawn by the Service architecture scan.
AttachmentA file added to a record or a comment.
Audit findingOne problem Productmind found when it checked a scan's draft specification, shown above the draft. Not the same as the audit log.
Audit logThe record of who created, changed and deleted what in your workspaces and organization.
Auto-applyA scan's draft becoming the specification with nobody reviewing it first. Only a composition, or the first specification of an entry that feature discovery created, can apply itself, when nobody edited the document during the scan and no finding blocks it.
Automated ingestionYour helpdesk sending each resolved or closed ticket to Productmind by itself, instead of you importing it.
AutomationA rule that runs an agent for you when something happens in your workspace, or on a schedule.
Automation-readyAn agent whose tools are safe to run with nobody watching, so an automation can run it.
Available to agentsA knowledge base document's switch that lets Productmind's AI search and use it.
Billing periodYour month, counted from the day and time of day your subscription started, not from the calendar. Allowances start again at the start of each one, also when you pay for a year ahead.
Blocking findingAn audit finding with a red blocking badge. It keeps a scan's draft from applying itself, but never stops you from applying it.
Blog postA long-form article for your customers about features a Marketing release is about, or passages of its published release notes. Productmind drafts it, your team edits it together, and you approve it in Marketing. A Marketing release can have any number of them.
Bug fix sessionA session that fixes bug reports, with a plan you approve first.
Bug reportA reported problem with your product, which you triage and can turn into a task. Its ID starts with BUG.
BuildRunning a session's approved plan, so the agent writes the code and opens pull requests, in the cloud or on your own machine.
Burst limitAn earlier name for a daily or weekly limit. The Usage tab now calls it a Limit.
Business rulesThe rules a component's code enforces, such as a limit or a condition, found by reading the code.
Cancel sessionThe session header's button (an X) that ends a coding session's run, once you confirm: the agent and any live preview stop, and the session becomes Cancelled. It keeps its plan, conversation, branch, pull requests, tasks and bug reports, and Draft a new plan starts it again.
CapA plan limit on how many of something you can have, such as workspaces, users or a workspace's features, or on how many referrals each person can send.
ChecksA pull request's CI results, as your Git provider reports them.
CitationThe link at the end of a statement in a specification to the code that supports it, at the commit the scan read.
Cloud allowanceThe Included cloud allowance: an amount in euros your plan includes each billing period for time in Productmind's cloud. Every sandbox size draws on it by the second at its own price per hour, and one overage switch decides what happens past it for every size.
Code pathA file or a folder on a code surface.
Code surfaceThe files and entry points that implement a feature in one of its linked applications, recorded by scans or added by you. Findings call it the surface.
Coding agentAn AI coding tool such as Claude Code, Codex or GitHub Copilot. The Productmind agent can drive Claude Code or Codex.
CommentA message on a record, where you can @mention people.
ComponentThe part of a feature that lives in your code, such as a screen, a service or a data store. Its ID starts with COMP.
CompositionA feature's specification written from its components' specifications, when you choose Compose from components.
Connected clientsThe browsers, devices and tools signed in to your account, listed in your profile, where you can revoke them.
ConnectionYour workspace's link to a tool on Integrations, such as Jira or Slack. Each workspace connects its own tools.
ContactA person your feedback comes from.
Custom roleA role made for your organization by Productmind, besides the built-in ones.
Customer emailA personal email to one customer about a Marketing release: what they asked for and the bugs they reported that it covers, then a short summary from its public release notes. Your team reviews it before sending, and every email sent is logged.
Customer meetingA conversation with a customer that you log, or that the notetaker records, and Productmind summarizes.
Customer release notesWhat you tell your customers in a Marketing release: a headline, a summary and what they can now do, as one document your team edits and comments on together. A Marketing release can have several, each published on its own. Also called public release notes. Written and published in Marketing.
Customer supportSupport conversations brought in from your helpdesk, so the requests and bugs in them reach your backlog.
Customized roleA built-in role whose permissions Productmind changed for your organization.
Daily limitThe Daily AI credit limit: the most AI credits agents and coding sessions can use together in a day, paid ones included. A day starts at your subscription's time of day. It stops work until the day resets, and overage and your Pay-as-you-go balance never lift it.
DashboardThe page that shows where your product work stands, from feedback to release.
DecisionA recorded reason for a change, kept in a feature's Decisions panel or a feature request's Decisions tab.
Decision for the product ownerSomething a scan found the code settles today that a person should confirm, listed in its draft with the rule it would change.
Deep readAn AI scan of one feature's or component's code that drafts its specification. It uses AI credits, and runs only when you start it.
Default sandboxThe sandbox a session runs in when it doesn't pick one. A workspace has at most one.
Development branchYour team's branch for new work. Releases merge it into the production branch.
Documented-but-falseA statement the code contradicts, in a specification, a comment or a message. A scan reports it rather than quietly dropping it.
DomainA named group of features in the feature catalog, such as Billing. It holds no code of its own. Its ID starts with DOM.
EndpointOne address of a running live preview, such as the web client or the API.
Engineering planA session's document: what the agent will change and how. You review it before anything is built.
Entry pointA way into a feature's code that a scan recorded on its code surface, such as a screen, an API route or a scheduled job.
FeatureA capability of your product, with a specification, a status and the tasks that build it. Its ID starts with FEAT.
Feature catalogThe tree of domains, features and components that describes what your product does.
Feature discoveryA scan that indexes all your repositories and builds your feature catalog from the code. When indexing can't build one, you can choose to read the code with AI instead.
Feature mapThe feature catalog drawn as a tree of cards, with the applications each feature spans.
Feature requestA request or idea for your product, before it becomes a feature. Its ID starts with PFR.
Feedback widgetA chat window you embed in your website, where visitors give feedback to an AI interviewer.
GrantOne Productmind Architect's temporary access to one session, created when an architect picks up your request.
HelpdeskYour support tool, Freshdesk or Zendesk, where support tickets live.
Included AI creditsThe AI credits your plan includes each billing period, for agents and coding together. Past them, extra use is billed at your plan's price per 1M AI credits while overage is on. What you don't use doesn't roll over.
IntegrationA connection between Productmind and another tool your team uses.
InterestEvidence that someone wants a feature request: a customer interest, stakeholder input or other input. On a bug report, a customer conversation the bug came up in.
Internal meetingA meeting your team has with itself, recorded and summarized.
InterviewOne conversation between the AI interviewer and one person, by email, by voice or in the feedback widget. Every interview belongs to a research.
InvitationAn email asking someone to join your organization. It works for 7 days.
Invoiced organizationAn organization that pays by invoice under an agreement with Productmind, instead of by card.
Just-in-time memberSomeone your identity provider signs in, who joins your organization automatically.
Knowledge baseThe documents, files and web pages that give Productmind's agents context about your product.
LimitA daily, weekly or monthly ceiling on the AI credits agents and coding sessions can use together, paid ones included. It is never billed, overage and your Pay-as-you-go balance never lift it, and it stops work until it resets.
LineageHow every piece of work connects, from the feedback that asked for it to the release that shipped it.
LinkThe connection between a feature and an application that implements it. It also holds the feature's code surface in that application.
link-feature-applicationThe API operation Link a feature to an application it depends on, which scans use to record a feature's code surface.
Live previewThe running application a session is changing, open in your browser.
Live sessionA session with no plan up front, where you change the running application by chatting with the agent while you watch its live preview.
Marketing materialsWhat Productmind writes for customers in a Marketing release: its public release notes and its blog posts. The Marketing list can show only the Marketing releases with materials, or only those without.
Marketing releaseOne announcement to your customers in Marketing, with its own title and a status you set (Planned, In progress, Announced or Archived). It is written from its sources, any mix of releases, features, feature requests and customer feedback, and holds its public release notes, blog posts and customer emails. Its ID starts with MKT.
MCPThe Model Context Protocol, the open standard AI assistants use to connect to Productmind.
MCP serverOn Integrations: a remote server whose tools your Productmind agents can use.
MCP Workflow SkillThe workflow Productmind gives a connected AI assistant when it works on your tasks.
MemberA person in your organization: active, invited or deactivated.
MentionYour name written with @ in a comment, a description or a document. It notifies you.
Monthly limitThe Monthly AI credit limit: the most AI credits agents and coding sessions can use together in a billing period, paid ones included. Past it, work stops until your next billing period starts, whatever the overage switch says and whatever your Pay-as-you-go balance holds. A plan without one has no monthly limit.
Multi-factor authenticationConfirming a code from an authenticator app when you sign in. Your profile calls it two-factor authentication.
NoteAn idea or observation on the notes wall, which you can convert into a feature request, a feature, a task or a bug report.
NotetakerProductmind's meeting bot, which joins a customer or internal meeting to record it.
Open questionSomething a scan couldn't determine from the code, listed in its draft with what it searched and what to check.
OrganizationYour company's account in Productmind. Its members, roles, sign-in rules and billing apply to all its workspaces.
OverageUsage beyond a monthly allowance, such as your included AI credits, paid from your Pay-as-you-go balance when you turn it on. It never lifts a daily, weekly or monthly limit.
OwnerThe one member with access to everything, including transferring ownership.
Pay as you goUsage your plan includes none of, charged per unit from the first one, at your plan's unit price. Plan comparisons show None, pay as you go in its place.
Pay-as-you-go balanceA prepaid balance, in euros, that your usage, overage and Productmind support are paid from. You top it up by card, or with iDEAL where Stripe offers it. Not the same as AI credits, which measure usage.
PinThe numbered marker of a comment on a live preview. Its color shows whether the comment is waiting, queued, being worked on or answered.
PipelineThe path work takes through Productmind, drawn on the Dashboard as cards in columns, from feedback on the left to releases on the right.
PlanWhat your organization subscribes to. It sets your limits.
Plan sessionA session for tasks: the agent drafts an engineering plan, you approve it, then the agent builds it.
pmThe command that runs the Productmind CLI.
PM agentThe chat in the activity panel, where you ask Productmind's AI about your workspace and follow coding sessions.
Preview commentA comment on an element or an area of a session's running live preview, which you send to the session's agent to fix.
Primary contactThe main person at an account, marked with a star.
Product overviewThe Overview system document of your knowledge base, which describes your product.
Production branchThe branch that holds what is live. Releases merge into it.
Productmind agentThe pm agent command, which builds sessions on your own machine.
Productmind ArchitectA Productmind engineer who reviews a coding session when your team asks for a code, architecture or product review. Shown as PM Architect.
Productmind CLIProductmind's command-line tool for your computer.
Productmind cloud sessionsWhether a workspace's coding sessions may run on the Productmind cloud. When it is off, they run only on the Productmind agent of the person who starts them.
Productmind support requestA request to Productmind for human help: a question, a bug report, a feature request, or a Productmind Architect's code, architecture or product review of a coding session. Its ID starts with SUP.
ProdutypingProductmind's name for production-ready prototyping in your applications: a Live session where you describe a change, watch it happen in the live preview, then open a pull request for review.
PromotionThe decision to build a feature request, which turns it into a feature or a task and links the two.
ProposalAn update to a knowledge base document or a feature specification that a coding session or a scan suggests, for you to approve before it applies. A scan's proposal is also called its draft, and some drafts apply themselves: see Audit findings.
Pull requestThe change a session proposes to your repository, for you to review and merge.
Queued messageA message you sent while the agent was working, which waits until it finishes its current step.
QuoteWhat a Productmind Architect sends when a review needs more hours than you approved: an estimate or an extension, for an Owner or Admin to approve.
ReceiptA line in a scan's draft that reports a check the scan ran. The Code surface receipt counts the files the draft was written from.
Recovery codeA one-time code that signs you in when you can't reach your authenticator app.
ReferralA personal sign-up link a current customer sends, so someone can start an organization of their own.
ReleaseA set of changes you merge, tag and announce together. Its ID starts with REL.
ResearchA set of agentic interviews with the same objectives and questions, and its own feedback widget. Your plan counts them as Studies per workspace.
Review mode (live preview)Commenting on what you see in a session's running application, then sending the comments to the agent, which fixes the code and replies under each one.
Review mode (specifications)Commenting on a feature's or a component's specification, then sending the comments to an agent that proposes a revision of the passages they point at.
Review taskThe task Productmind creates when you ask a teammate to review a session's pull requests.
RoleWhat a member can do. Every member has one role in the organization.
RunOne time an automation's agent ran, with its status, its input and its output.
Run-as userThe person an automation acts as: the one who created it.
RunnerThe machine in Productmind's cloud that runs one coding session, its live preview or one scan step. It serves only that session, never another session or another organization.
SandboxWhere and how a coding session runs your application. Its ID starts with SBX.
Sandbox hourThe unit hosted previews and hosted coding sessions are counted in, per sandbox size, measured by the second. Each size has its own price per hour, at which its hours draw on the cloud allowance, and its own sandbox hours limit. Runs on your own machine are free.
Sandbox hours limitThe most hours one sandbox size can run in Productmind's cloud in a billing period. A ceiling no overage lifts: at zero, the size isn't included.
Sandbox sizeThe machine size in Productmind's cloud that a hosted preview or a hosted coding session is scheduled on: Small (2 vCPU, 8 GB memory, 50 GB storage), Medium (4 vCPU, 16 GB, 60 GB, the default) or Large (8 vCPU, 32 GB, 100 GB). You can pick any of them on every plan, and your plan sets the price of each and can cap or leave out each one.
ScanAn AI agent reading your repositories to document them, find their features or rate them. Scans never change your code.
Scan estimateWhat a scan of your repositories is expected to cost, as a range in AI credits and euros, or, for a scan on your own machine, as its per-session price. A scan starts only when your Pay-as-you-go balance covers the lower end.
Scan summaryWhat a scan says it did, shown under Review proposed changes when you review its draft.
SeatA place for one person on your plan. Active and invited members each take one.
ServiceSomething that runs inside an application's repository, such as a web front end or an API. A service with a port is what a live preview opens.
SessionA coding session: an AI coding agent works on your tasks or bug reports in your repositories and opens pull requests. Its ID starts with SSID.
SeverityHow bad a bug is: Minor, Major, Critical or Blocker.
Single sign-onSigning in to Productmind through your company's identity provider.
SkillReusable instructions your agents and coding sessions can apply.
SpecificationA feature's document: what it does, for whom, and the rules it follows. Your team writes it together, with AI or from your code.
StageHow far a customer account is in its relationship with you, from Prospect to Churned.
Stakeholder inputAn interest recorded for someone on your team, with their role and their position.
Stop previewThe square button beside a running live preview. It stops the preview, and the agent's current step if it is working. The session stays open, and Start preview brings the preview back.
Stop turnThe square button in a session's composer while the agent works. It stops the agent's current step, not the session: your next message carries on from there.
Strategic accountAn account you've marked as especially important, shown with a star.
SubtaskA task that sits under another task. Any task can be one, whatever its type: Subtask is also a type label you can give a task, and it doesn't decide where the task sits.
SummaryWhat Productmind writes from a conversation or a meeting. Summarizing a customer conversation also creates feature requests and bug reports.
Support accessYour control over whether Productmind Architects may open your sessions for a review.
Support caseA support conversation kept in Productmind, imported from your helpdesk or added by hand.
Support centerThe tab in the activity panel where you ask Productmind's support assistant how to use Productmind and its API, and open support requests to Productmind. The assistant has no access to your workspace data.
SurfaceShort for code surface: the files and entry points recorded for a feature in one application.
System documentOne of the four documents every knowledge base starts with: Overview, User experience, Competitors and Personas.
TagA short label you apply to records to group and filter them, saved in lowercase with hyphens between words.
TaskA concrete piece of engineering work, with a status, a priority and assignees. Its ID starts with TASK.
TeamA group of members you can assign work to at once.
Team roleA role a team gives its members on top of their own. It adds permissions, never removes them.
TemplateThe shape of a document an agent writes, or the instructions for a job it does.
TicketA conversation's record in your helpdesk. Importing a ticket creates a support case.
TranscriptThe written record of a call or a ticket, on a meeting's Meeting notes tab or a support case's Case notes tab.
TriggerWhat starts an automation: a workspace event, a schedule, or only by hand.
Two-factor authenticationWhat your profile calls multi-factor authentication.
UnknownThe account for contacts that don't belong to any company.
User-hosted sessionA coding session the Productmind agent runs on your own machine, on your own Claude Code or Codex account. It uses no AI credits. If your plan charges for them, each one is charged once, at your plan's per-session rate, the first time it runs, scans included.
Weekly limitThe Weekly AI credit limit: the most AI credits agents and coding sessions can use together in a week, paid ones included. A week is always 7 days, counted from the start of your subscription, and it runs on from one billing period into the next. It stops work until the week resets, and overage and your Pay-as-you-go balance never lift it.
Welcome messageThe first message a visitor sees when they open the feedback widget.
Widget keyThe code that ties a feedback widget to its research.
Working branchThe branch a session's agent commits to, cut from each repository's base branch.
WorkspaceA separate space for one product, inside your organization.
Workspace accessWhich workspaces a member can use. Owners and Admins get the workspaces that exist when they join. Everyone else uses the workspaces they are given.
Workspace adminA member who manages one workspace (its name, who can open it, resetting and deleting it) without being an Owner or Admin of the organization.
Workspace settingsWhere you manage the workspace you're in: General, with its Danger zone, and Permissions. You open it from the workspace switcher.
Last modified on