PrompteryxBeta
    ⌘K
    PlatformSupport
    Menu
    IntroductionInstall Desktop & extensionQuick StartPlatform Overview
    OverviewWhat you can makeAI Browser Agent in the chatModels and AutoConnectorsSchedulesProjectsExplore and publishingShare linksMemory and SkillsPersonal agent (beta)Welcome, settings and skinsPlans, credits and top-upsPrompteryx DesktopGroups (pilot)Phone app (early access)
    OverviewCreating WorkflowsTesting & DebuggingScheduling
    OverviewGetting StartedCommandsRecordingSettingsStructured OutputParallel AgentsTips & Best Practices
    OverviewInstallationSidebar AgentUsing Attached FilesExecute on Current TabExporting to Visual Studio
    Overview
    OverviewOpen PageClickClick CoordinatesFillExtract TextScreenshotWaitPress KeyFocus ElementBlur ElementSelect TextHighlight ElementSet ViewportGet Bounding BoxScrollHoverSelect OptionAssertionsCookiesEvaluate JavaScriptExample Workflows
    OverviewPrompteryx Browser AgentOpenAI
    OverviewSchedule TriggerWebhook Trigger
    OverviewGoogle SheetsGoogle CalendarConnect Hub (3,000+ apps)Example Workflows
    AI Vision SettingsChrome Automation ProfilesCloud BrowserBrowser Profiles & Persistent ContextVideo RecordingExecution LogsSharing & CollaborationEdit History
    Selectors GuideVariables
    Start hereGet an API keyRun the AI browser agentDrive a cloud browserRun workflows from anywhereUse connected apps (Connect Hub)SDK (TypeScript)CLIMCP for AI assistants
    REST APIWebhooksMCP toolsCloud Browser APICloud Browser quickstartMigrate from Browserbasen8nZapierMake.com

    Need help? Contact Support

    1. Documentation
    2. App
    3. Memory And Skills

    Memory and Skills

    Memory keeps track of what you tell it, so you don't repeat yourself. Skills are instructions for work you do again and again.

    Memory

    Open Memory from your account menu (bottom left of the sidebar) to see what the assistant remembers about you and to control it. Memories are background facts, never instructions.

    The Memory page: How memory works with switches for using saved memories, learning from chats, referencing past chats and sensitive topics, then the About you section
    Memory: the switches, then About you and your saved memories further down.
    Memory on
    The main switch. Off means nothing is remembered or used.
    Use saved memories
    The assistant uses what it remembers to tailor answers.
    Learn from your chats
    It notices lasting things you mention, such as your role, preferences or ongoing work, and saves them. Each time, you see a “Memory updated” note with Undo. Saying “remember…” always works.
    Reference past chats
    It can see the titles of your recent chats and search your past conversations when you refer back to them.
    Allow sensitive topics
    Off by default: health, religion, politics, ethnicity, sexuality and similar topics are not learned automatically.
    • Passwords, card, bank and ID numbers are never stored.
    • Each saved memory can be edited, pinned (always used), kept but not used in chats, or deleted, and you can see its earlier wordings.
    • Projects keep their own memories: what is learned in a project's chats stays in that project.

    Memory in Settings

    Settings → Memory puts the same controls in one place, with a few more:

    • Search and reference chats, Generate memory from chats and Include sensitive topics in memory: the same switches as on the Memory page.
    • Import memory from other AI providers: bring what ChatGPT, Claude or Gemini know about you. You get a prompt to paste there, paste its answer back, and choose what to keep.
    • Export memory: download everything as a .zip of Markdown files.
    • Everything it remembers, grouped (You, Topics, Areas and Projects). Open a group to see or delete single memories.
    • Tell Prompteryx what to change, for example "I've moved to Leeds" or "forget that I use Windows". You see the changes it suggests and untick any you don't want before they are saved. This uses a few AI credits.
    About you

    Your own instructions for every chat: what to call you, and what it should know about you (your work, where you are, what you use it for).

    Turn Use in chats off to pause it. It is used in incognito chats too.

    Incognito and memory off

    The eye icon in the message box starts an incognito chat: it isn't saved, and it never uses or creates memories.

    To keep a chat but leave memory out of it, turn memory off for that chat from the message box instead.

    Skills

    A skill is a set of reusable instructions, such as how you like meeting notes or code reviews done. The assistant applies a skill automatically when a request matches it, or when you type its name with a slash, such as /meeting-notes. You can also pick one with the lightning icon in the message box.

    The Skills page: New skill, Write with AI, Upload SKILL.md and Create in chat buttons, a search box, and cards for built-in skills such as Animated video, Code review, Data analysis and Email writer
    Skills: built-in skills to start with, and your own.
    Make your own
    • New skill: write it yourself.
    • Write with AI: describe it and let the assistant draft it.
    • Upload SKILL.md: bring a skill file you already have.
    • Create in chat: build it in a conversation.
    Built-in and automatic

    Built-in skills (for example Code review, Data analysis, Email writer and Meeting notes) are ready to use. Switch any skill off with its toggle.

    Automatic skills apply when a request matches; Only when called skills wait for /name.

    Open Skills from your account menu for the page above. Settings → Skills is a directory: Yours lists the skills you made or switched on, Discover lists ready-made ones to add. Search them, filter by category, press Try to use one in a new chat, or Add to make your own.

    Ready-made skills in Discover

    Discover holds skills written by Prompteryx and open-source skills made for Claude: from Anthropic (Apache 2.0 licence) and from Alireza Rezvani's claude-skills (MIT licence), such as copywriting, CFO and CEO advisors, pricing strategy and weekly review.

    Open-source skills are used as their authors wrote them, with a short note added so they work with Prompteryx's tools. Each card says who wrote it and under which licence, and its details link to the original.

    A skill only appears once it has passed a test run on the real assistant. Beta means part of it could not be tested, for example because it needs your own email. Skills that need things Prompteryx chats don't have, such as a terminal or your own code repository, are not included.

    The personal agent has its own memory and identity file, separate from these.