[{"data":1,"prerenderedAt":880},["ShallowReactive",2],{"help-category-\u002Fgetting-started\u002Fglossary":3,"help-article-\u002Fgetting-started\u002Fglossary":4,"related-articles-\u002Fgetting-started\u002Fglossary":339},[],{"id":5,"title":6,"body":7,"category":319,"description":320,"draft":321,"extension":322,"meta":323,"navigation":324,"order":325,"path":326,"relatedArticles":327,"seo":333,"slug":334,"stem":335,"updatedAt":336,"__hash__":337,"excerpt":320,"searchText":338},"help\u002Fhelp\u002Fgetting-started\u002F12.glossary.md","Glossary",{"type":8,"value":9,"toc":283},"minimark",[10,14,17,22,25,29,37,41,44,48,51,55,65,69,76,80,83,87,90,94,97,101,104,108,111,115,118,122,125,129,135,139,144,148,151,155,158,162,165,169,172,176,179,183,189,193,196,200,207,210,213,217,228,231,239,243,250,253,263,266,269,273,276,280],[11,12,13],"p",{},"This glossary covers the key terms used across MultiClaw and its documentation. If you hit an unfamiliar word in another article, start here.",[15,16],"hr",{},[18,19,21],"h3",{"id":20},"agent","Agent",[11,23,24],{},"An AI assistant you configure to perform tasks on your behalf. Each agent has a name, a personality (system prompt), an AI model, and a set of tools it can use.",[18,26,28],{"id":27},"ai-model","AI model",[11,30,31,32,36],{},"The language model (for example, from OpenAI or Anthropic) that powers an agent's reasoning and actions. You choose a model when you create an agent and can change it later from the agent's settings. Workspace guardrails can restrict which model providers are available. See also: ",[33,34,35],"strong",{},"Token",".",[18,38,40],{"id":39},"approval","Approval",[11,42,43],{},"Your explicit confirmation that an agent's proposed plan is ready to execute. When an agent presents a plan during a conversation, you can review and confirm before the agent proceeds — or redirect the agent with further instructions.",[18,45,47],{"id":46},"audit-trail","Audit trail",[11,49,50],{},"A time-stamped record of agent actions, tool calls, and conversation events. Useful when you need to review what an agent did and when.",[18,52,54],{"id":53},"browser-automation","Browser automation",[11,56,57,58,61,62,36],{},"Agents can control a real web browser to complete web-based tasks on your behalf: navigating pages, clicking buttons, and filling forms. When using a cloud desktop, the browser runs inside the remote virtual machine, not on your local machine. See also: ",[33,59,60],{},"Recording",", ",[33,63,64],{},"Skill",[18,66,68],{"id":67},"channel","Channel",[11,70,71,72,75],{},"An application or platform an agent can interact with. The currently supported channel is ",[33,73,74],{},"Slack",", which lets agents respond to direct messages and channel mentions.",[18,77,79],{"id":78},"cloud-desktop","Cloud desktop",[11,81,82],{},"A remote virtual machine managed by MultiClaw where agents can run continuously, independent of your local machine. Also called an \"instance\" in some parts of the interface.",[18,84,86],{"id":85},"connected","Connected",[11,88,89],{},"The state when the desktop app is signed in to MultiClaw Cloud and the OpenClaw gateway is running. Agents can only run tasks when connected.",[18,91,93],{"id":92},"conversation","Conversation",[11,95,96],{},"A chat thread between you and an agent. Contains your messages and the agent's responses, including any tool calls the agent makes along the way.",[18,98,100],{"id":99},"dashboard","Dashboard",[11,102,103],{},"The workspace home view in MultiClaw Cloud. Provides navigation to your workspace's agents, instances, skills, users, usage, and guardrails.",[18,105,107],{"id":106},"desktop-app","Desktop app",[11,109,110],{},"The MultiClaw application installed on your Mac, Windows, or Linux machine. The primary interface for chatting with agents and managing your workspace.",[18,112,114],{"id":113},"gateway","Gateway",[11,116,117],{},"The local web server that OpenClaw runs on your machine. The desktop app connects to it to communicate with agents. If the gateway is not running, agents cannot respond.",[18,119,121],{"id":120},"guardrails","Guardrails",[11,123,124],{},"Rules set by a workspace owner that restrict what agents can do, applying across all workspace members. Examples: limit which AI models agents may use, block specific tools or categories of tools, require human approval for certain action types, or cap the token budget per task.",[18,126,128],{"id":127},"instance","Instance",[11,130,131,132,134],{},"See ",[33,133,79],{},". Used interchangeably in some parts of the interface.",[18,136,138],{"id":137},"local-agent","Local agent",[11,140,141,142,36],{},"An agent stored and running on your own machine rather than in MultiClaw Cloud. Local agents are only accessible on the device where they are stored, do not sync across devices, and only run while your machine is on. See also: ",[33,143,79],{},[18,145,147],{"id":146},"mcp-model-context-protocol","MCP (Model Context Protocol)",[11,149,150],{},"An open protocol that lets agents connect to external tool servers. Each MCP server exposes a set of tools agents can call.",[18,152,154],{"id":153},"mcp-server","MCP server",[11,156,157],{},"A server that implements the Model Context Protocol and provides agents with access to specific tools or data sources (for example, GitHub, Google Drive, or a database).",[18,159,161],{"id":160},"multiclaw-chrome-extension","MultiClaw Chrome Extension",[11,163,164],{},"A browser extension (Chrome, Edge, Brave, and other Chromium-based browsers) used to record browser actions so you can teach agents new workflows and save them as skills.",[18,166,168],{"id":167},"multiclaw-cloud","MultiClaw Cloud",[11,170,171],{},"The cloud platform at multiclaw.ai that manages your account, workspace, team, and cloud desktops.",[18,173,175],{"id":174},"openclaw","OpenClaw",[11,177,178],{},"The open-source AI agent engine that powers MultiClaw. Runs locally or on a cloud desktop and stores all agent data as plain files. Managed automatically by the desktop app.",[18,180,182],{"id":181},"personality","Personality",[11,184,185,186,188],{},"The system prompt that defines how an agent thinks, responds, and behaves. Written in plain English and set from the ",[33,187,182],{}," tab on the agent's profile. A focused, role-specific personality produces more consistent results than a generic one.",[18,190,192],{"id":191},"plan","Plan",[11,194,195],{},"An ordered list of steps an agent proposes within a conversation before carrying out a task. You can review and redirect the plan before the agent acts.",[18,197,199],{"id":198},"quick-chat","Quick chat",[11,201,202,203,206],{},"A conversational interaction where the agent answers questions, brainstorms, or summarizes rather than taking action on your behalf. Contrast with a ",[33,204,205],{},"task",", where the agent drafts a plan and executes steps. Both happen inside a conversation — the difference is whether you ask a question or tell the agent to do something.",[18,208,60],{"id":209},"recording",[11,211,212],{},"A captured sequence of browser actions created with the MultiClaw Chrome Extension. You can convert a recording into a reusable skill.",[18,214,216],{"id":215},"role","Role",[11,218,219,220,223,224,227],{},"Each workspace member has a role that controls what they can see and do. Two roles exist: ",[33,221,222],{},"Owner"," and ",[33,225,226],{},"Member",". The Owner created the workspace and can invite or remove members. Members can use workspace resources but cannot manage membership.",[18,229,64],{"id":230},"skill",[11,232,233,234,238],{},"A reusable procedure that an agent can follow, created from a browser recording or written manually. Stored as a ",[235,236,237],"code",{},"SKILL.md"," file in OpenClaw.",[18,240,242],{"id":241},"subagent","Subagent",[11,244,245,246,249],{},"An agent that another agent can delegate subtasks to. When subagents are configured, the parent agent can break a large task into smaller pieces and propose which parts to hand off. This delegation appears in the plan you approve before execution begins. You can configure subagents from the ",[33,247,248],{},"Subagents"," tab on the agent's profile.",[18,251,252],{"id":205},"Task",[11,254,255,256,258,259,61,261,36],{},"A unit of work that an agent carries out during a conversation. When you ask an agent to act, the agent drafts a ",[33,257,191],{}," outlining the proposed steps. After you review and approve the plan, the agent executes it on your local machine or a cloud desktop, depending on the agent type. Results stream back to your conversation in real time. See also: ",[33,260,192],{},[33,262,40],{},[18,264,35],{"id":265},"token",[11,267,268],{},"The basic unit of text that AI language models process. Guardrails let workspace owners set a maximum token budget per task, which helps limit costs from long-running or looping agents.",[18,270,272],{"id":271},"workflow","Workflow",[11,274,275],{},"A recorded or scripted sequence of steps that an agent can follow. Typically created from a browser recording and converted into a skill in the desktop app.",[18,277,279],{"id":278},"workspace","Workspace",[11,281,282],{},"The top-level organizational unit in MultiClaw Cloud. All agents, team members, and cloud desktops belong to a workspace.",{"title":284,"searchDepth":285,"depth":285,"links":286},"",2,[287,289,290,291,292,293,294,295,296,297,298,299,300,301,302,303,304,305,306,307,308,309,310,311,312,313,314,315,316,317,318],{"id":20,"depth":288,"text":21},3,{"id":27,"depth":288,"text":28},{"id":39,"depth":288,"text":40},{"id":46,"depth":288,"text":47},{"id":53,"depth":288,"text":54},{"id":67,"depth":288,"text":68},{"id":78,"depth":288,"text":79},{"id":85,"depth":288,"text":86},{"id":92,"depth":288,"text":93},{"id":99,"depth":288,"text":100},{"id":106,"depth":288,"text":107},{"id":113,"depth":288,"text":114},{"id":120,"depth":288,"text":121},{"id":127,"depth":288,"text":128},{"id":137,"depth":288,"text":138},{"id":146,"depth":288,"text":147},{"id":153,"depth":288,"text":154},{"id":160,"depth":288,"text":161},{"id":167,"depth":288,"text":168},{"id":174,"depth":288,"text":175},{"id":181,"depth":288,"text":182},{"id":191,"depth":288,"text":192},{"id":198,"depth":288,"text":199},{"id":209,"depth":288,"text":60},{"id":215,"depth":288,"text":216},{"id":230,"depth":288,"text":64},{"id":241,"depth":288,"text":242},{"id":205,"depth":288,"text":252},{"id":265,"depth":288,"text":35},{"id":271,"depth":288,"text":272},{"id":278,"depth":288,"text":279},"getting-started","Quick-reference definitions for the key terms used across MultiClaw and its documentation.",false,"md",{},true,12,"\u002Fgetting-started\u002Fglossary",[328,329,330,331,332],"getting-started\u002Fwhat-is-multiclaw","getting-started\u002Fhow-multiclaw-works","agents\u002Fwhat-are-agents","skills-workflows\u002Fwhat-are-skills","tools-integrations\u002Fwhat-is-mcp",{"title":6,"description":320},"glossary","help\u002Fgetting-started\u002F12.glossary",null,"5eQUnVQnPzRJmqSr0W0V3envNSPjUY_C1gK3cFI1BZA","Glossary Quick-reference definitions for the key terms used across MultiClaw and its documentation.",[340,507,729],{"id":341,"title":342,"body":343,"category":319,"description":494,"draft":321,"extension":322,"meta":495,"navigation":324,"order":497,"path":498,"relatedArticles":499,"seo":501,"slug":502,"stem":503,"updatedAt":504,"__hash__":505,"excerpt":494,"searchText":506},"help\u002Fhelp\u002Fgetting-started\u002F01.what-is-multiclaw.md","What is MultiClaw?",{"type":8,"value":344,"toc":489},[345,361,364,369,372,375,415,418,422,425,428,437,440,444,447,474,481],[11,346,347,350,351,354,355,357,358,36],{},[33,348,349],{},"MultiClaw"," is a desktop app that lets you run ",[33,352,353],{},"AI agents",": software that does real computer work on your behalf. Instead of spending hours on repetitive work like research, data entry, and form-filling, you describe what you need in a ",[33,356,92],{},", and an agent handles the steps on your local machine or a ",[33,359,360],{},"cloud desktop",[11,362,363],{},"You talk to agents the way you'd talk to a colleague. Describe what you need, then check in when you want to.",[365,366,368],"h2",{"id":367},"what-multiclaw-does","What MultiClaw does",[11,370,371],{},"Most computer work follows a pattern: open a browser, navigate to a page, copy some data, paste it somewhere else, repeat. MultiClaw gives you a way to hand that work to an agent while you focus on what matters.",[11,373,374],{},"The process works like this:",[376,377,378,385,391,397,403],"ul",{},[379,380,381,384],"li",{},[33,382,383],{},"You describe what you need"," in plain language, no code required. Start a conversation, explain what you're after, and the agent gets to work.",[379,386,387,390],{},[33,388,389],{},"The agent proposes an approach"," right in the conversation. You can review it, ask questions, or redirect before the agent acts.",[379,392,393,396],{},[33,394,395],{},"Work runs locally or in the cloud."," Choose your own machine or a cloud desktop in your workspace.",[379,398,399,402],{},[33,400,401],{},"Agents use browsers and web-based tools",", with additional capabilities available through MCP servers and Slack integration.",[379,404,405,408,409,412,413,36],{},[33,406,407],{},"You stay informed",": every agent action is logged in an ",[33,410,411],{},"audit trail"," you can review in ",[33,414,168],{},[11,416,417],{},"You can run multiple agents at the same time and manage them all from the desktop app.",[365,419,421],{"id":420},"who-its-for","Who it's for",[11,423,424],{},"If you spend hours each week on repetitive computer work (copying data between systems, filling out forms, researching across dozens of tabs, sending the same outreach messages), MultiClaw is built for you.",[11,426,427],{},"Without MultiClaw, you do those steps yourself: click, copy, paste, switch tabs, repeat. With MultiClaw, you describe the outcome you want, and an agent follows the steps on your behalf. You stay in control of the process and review results as the agent works.",[11,429,430,431,433,434,436],{},"MultiClaw works for individuals and teams. Whether you're working solo or managing agents in a ",[33,432,278],{},", there are built-in governance controls, audit trails, and workspace-level ",[33,435,120],{}," (rules that define what agents can and cannot do) to maintain accountability.",[11,438,439],{},"You don't need a technical background. If you can describe what you want done, you can use MultiClaw.",[365,441,443],{"id":442},"how-it-works","How it works",[11,445,446],{},"MultiClaw is made up of four components that work together. You use the desktop app daily; the rest run in the background.",[376,448,449,454,459,469],{},[379,450,451,453],{},[33,452,107],{},": the application on your Mac, Windows, or Linux machine where you chat with agents and manage your work.",[379,455,456,458],{},[33,457,168],{},": the web platform that manages your account, workspace, team, and cloud desktops.",[379,460,461,464,465,468],{},[33,462,463],{},"Chrome Extension",": records your browser actions so you can save them as reusable ",[33,466,467],{},"skills"," that agents can follow.",[379,470,471,473],{},[33,472,175],{},": the open-source agent engine that runs locally or on a cloud desktop, depending on your setup.",[475,476,478],"callout",{"type":477},"tip",[11,479,480],{},"You don't need to understand every component to get started. The desktop app is your main interface. Start there.",[11,482,483,484,36],{},"For a detailed look at each component and which ones are required, see ",[485,486,488],"a",{"href":487},"\u002Fgetting-started\u002Funderstand-the-multiclaw-products","What MultiClaw includes",{"title":284,"searchDepth":285,"depth":285,"links":490},[491,492,493],{"id":367,"depth":285,"text":368},{"id":420,"depth":285,"text":421},{"id":442,"depth":285,"text":443},"MultiClaw is a desktop app for running AI agents that do real computer work — on your machine or a cloud desktop.",{"icon":496},"rocket",1,"\u002Fgetting-started\u002Fwhat-is-multiclaw",[329,500],"getting-started\u002Funderstand-the-multiclaw-products",{"title":342,"description":494},"what-is-multiclaw","help\u002Fgetting-started\u002F01.what-is-multiclaw","2026-03-30","UxVDafPwH5JmMxQFobvD-x4JPqhxyrvePXkUrM_QkkQ","What is MultiClaw? MultiClaw is a desktop app for running AI agents that do real computer work — on your machine or a cloud desktop.",{"id":508,"title":509,"body":510,"category":319,"description":718,"draft":321,"extension":322,"meta":719,"navigation":324,"order":285,"path":720,"relatedArticles":721,"seo":724,"slug":725,"stem":726,"updatedAt":336,"__hash__":727,"excerpt":718,"searchText":728},"help\u002Fhelp\u002Fgetting-started\u002F02.how-multiclaw-works.md","How MultiClaw works",{"type":8,"value":511,"toc":708},[512,528,535,539,541,544,546,551,553,562,564,572,576,579,609,615,619,690,703],[11,513,514,515,518,519,521,522,524,525,527],{},"MultiClaw has four components: the ",[33,516,517],{},"desktop app"," you install on your computer, ",[33,520,168],{}," (the cloud platform at multiclaw.ai), the ",[33,523,161],{}," (an optional browser extension for recording workflows), and ",[33,526,175],{}," (the open-source agent engine that processes your requests and runs tasks). You interact with the desktop app daily — the rest work in the background. Understanding how they connect helps you know where your data lives and what to check when something goes wrong.",[11,529,530],{},[531,532],"img",{"alt":533,"src":534},"MultiClaw architecture — how the four components connect","\u002Fimages\u002Fmulticlaw-architecture-diagram.png",[365,536,538],{"id":537},"the-four-components","The four components",[18,540,107],{"id":106},[11,542,543],{},"The desktop app is the application you install on your Mac, Windows, or Linux computer. It's where you chat with agents, review plans, approve task execution, and monitor what's happening. The app connects to OpenClaw locally and to MultiClaw Cloud over the internet. It doesn't have its own database — it reads and writes OpenClaw's files directly.",[18,545,168],{"id":167},[11,547,548,550],{},[33,549,168],{}," is the cloud platform that manages your account, workspace, team membership, and cloud desktops. It also runs the relay server that routes signals between your desktop app and any agents running on cloud desktops. When you sign in, MultiClaw Cloud verifies your identity and issues the credentials everything else uses.",[18,552,161],{"id":160},[11,554,555,556,558,559,561],{},"The ",[33,557,161],{}," is an optional browser extension for Chrome, Chromium, Edge, Brave, and Arc. It records your browser actions — clicks, form fills, and page navigation — so you can save them as reusable ",[33,560,467],{}," that agents can follow. It communicates only with your local desktop app. You only need it if you want to record browser workflows.",[18,563,175],{"id":174},[11,565,566,568,569,571],{},[33,567,175],{}," is the open-source engine that processes your requests and carries out tasks. It runs either locally on your machine or on a ",[33,570,360],{}," — a remote computer managed by MultiClaw Cloud. OpenClaw stores your conversations, agent configurations, skills, and scheduled workflows as files, not in a proprietary database. MultiClaw handles all communication with OpenClaw; you never need to interact with it directly.",[365,573,575],{"id":574},"how-a-task-works","How a task works",[11,577,578],{},"From your perspective, running a task looks like a conversation. Here's what happens at each step:",[580,581,582,585,588,591,597,603,606],"ol",{},[379,583,584],{},"You type a message in the desktop app.",[379,586,587],{},"The desktop app sends your message to OpenClaw over a local connection.",[379,589,590],{},"OpenClaw processes your request using your configured AI model.",[379,592,593,594,596],{},"The agent drafts a ",[33,595,191],{}," — an ordered list of steps it proposes to take.",[379,598,599,600,36],{},"The desktop app shows you the plan. You review it, edit any steps if needed, and click ",[33,601,602],{},"Approve",[379,604,605],{},"The agent executes the plan — on a cloud desktop or your local machine, depending on the task.",[379,607,608],{},"Results stream back to your desktop app in real time as work completes.",[475,610,612],{"type":611},"note",[11,613,614],{},"By default, your approval is required before the agent takes any action. The agent cannot proceed past the plan stage without it. Scheduled tasks and automated workflows may run without real-time approval when explicitly configured.",[365,616,618],{"id":617},"what-stays-local-what-goes-to-the-cloud","What stays local, what goes to the cloud",[620,621,622,635],"table",{},[623,624,625],"thead",{},[626,627,628,632],"tr",{},[629,630,631],"th",{},"Data",[629,633,634],{},"Where it lives",[636,637,638,647,655,662,670,682],"tbody",{},[626,639,640,644],{},[641,642,643],"td",{},"Conversation history",[641,645,646],{},"On your machine (local mode) or on your cloud desktop (cloud mode)",[626,648,649,652],{},[641,650,651],{},"Agent configurations and skills",[641,653,654],{},"On whichever machine OpenClaw is running",[626,656,657,660],{},[641,658,659],{},"Your account, workspace, and team settings",[641,661,168],{},[626,663,664,667],{},[641,665,666],{},"Task signals and approvals",[641,668,669],{},"Routed through MultiClaw Cloud — your machine and cloud desktop do not connect directly",[626,671,672,675],{},[641,673,674],{},"Message content sent to AI models",[641,676,677,678],{},"Your configured AI model provider (e.g. OpenAI, Anthropic, Google) — see our ",[485,679,681],{"href":680},"\u002Flegal-compliance\u002Fprivacy-policy","Privacy Policy",[626,683,684,687],{},[641,685,686],{},"Files created during task execution",[641,688,689],{},"On the cloud desktop while the task runs",[475,691,692],{"type":477},[11,693,694,695,698,699,702],{},"In local mode, all OpenClaw data is stored in a standard folder on your machine (",[235,696,697],{},"~\u002F.openclaw\u002F"," on Mac and Linux, ",[235,700,701],{},"%USERPROFILE%\\.openclaw"," on Windows). You can back it up like any other folder.",[475,704,705],{"type":611},[11,706,707],{},"When you send a message, its content is transmitted to your configured AI model provider (such as OpenAI, Anthropic, or Google) for processing. Those providers handle this data under their own terms of service and privacy policies.",{"title":284,"searchDepth":285,"depth":285,"links":709},[710,716,717],{"id":537,"depth":285,"text":538,"children":711},[712,713,714,715],{"id":106,"depth":288,"text":107},{"id":167,"depth":288,"text":168},{"id":160,"depth":288,"text":161},{"id":174,"depth":288,"text":175},{"id":574,"depth":285,"text":575},{"id":617,"depth":285,"text":618},"MultiClaw has four components: a desktop app, a cloud platform, a browser extension, and OpenClaw. Here is how they work together.",{},"\u002Fgetting-started\u002Fhow-multiclaw-works",[328,722,723],"getting-started\u002Finstall-multiclaw","getting-started\u002Fstart-your-first-conversation",{"title":509,"description":718},"how-multiclaw-works","help\u002Fgetting-started\u002F02.how-multiclaw-works","WaNLNzjKyU019RKVKObaSoVhKfGZnfb0YlOMSUrycsE","How MultiClaw works MultiClaw has four components: a desktop app, a cloud platform, a browser extension, and OpenClaw. Here is how they work together.",{"id":730,"title":731,"body":732,"category":867,"description":868,"draft":321,"extension":322,"meta":869,"navigation":324,"order":497,"path":870,"relatedArticles":871,"seo":875,"slug":876,"stem":877,"updatedAt":336,"__hash__":878,"excerpt":868,"searchText":879},"help\u002Fhelp\u002Fagents\u002F01.what-are-agents.md","What are agents?",{"type":8,"value":733,"toc":862},[734,737,740,744,768,773,777,846,849,853,859],[11,735,736],{},"An agent is an AI assistant you configure to work for you. You give it a name, a personality, an AI model, and tools — then it can chat with you, draft plans, execute tasks, and repeat workflows automatically.",[11,738,739],{},"Before executing a task, an agent proposes a step-by-step plan for your review. You approve or reject the plan before any action is taken.",[365,741,743],{"id":742},"what-agents-can-do","What agents can do",[376,745,746,749,752,755,758,761],{},[379,747,748],{},"Answer questions and hold conversations",[379,750,751],{},"Draft step-by-step plans for tasks and wait for your approval",[379,753,754],{},"Execute tasks on a cloud desktop — browsing the web, filling forms, editing files",[379,756,757],{},"Use MCP tools to connect to apps like GitHub, Slack, Google Drive, and others",[379,759,760],{},"Run recorded browser workflows and reusable skills — each run is subject to your approval before execution",[379,762,763,764,767],{},"Delegate subtasks to other agents (",[33,765,766],{},"subagents",") — delegation is included in the plan you approve before execution begins",[475,769,770],{"type":611},[11,771,772],{},"If you configure agents to communicate with third parties — for example, sending emails or submitting forms on your behalf — you are responsible for ensuring those parties are informed they may be interacting with an automated AI system, as required by applicable law in your jurisdiction.",[365,774,776],{"id":775},"local-agents-vs-cloud-agents","Local agents vs cloud agents",[620,778,779,790],{},[623,780,781],{},[626,782,783,785,787],{},[629,784],{},[629,786,138],{},[629,788,789],{},"Cloud agent",[636,791,792,807,820,833],{},[626,793,794,799,805],{},[641,795,796],{},[33,797,798],{},"Where stored",[641,800,801,802,804],{},"Your machine (",[235,803,697],{},")",[641,806,168],{},[626,808,809,814,817],{},[641,810,811],{},[33,812,813],{},"Who can access it",[641,815,816],{},"Only you, on that device",[641,818,819],{},"Any workspace member with permission",[626,821,822,827,830],{},[641,823,824],{},[33,825,826],{},"Syncs across devices",[641,828,829],{},"No",[641,831,832],{},"Yes",[626,834,835,840,843],{},[641,836,837],{},[33,838,839],{},"Runs 24\u002F7",[641,841,842],{},"Only while your machine is on",[641,844,845],{},"Yes, on a cloud desktop",[11,847,848],{},"Use local agents for personal work you want to keep private. Use cloud agents when your team needs shared access or you want agents running continuously.",[365,850,852],{"id":851},"agent-personalities","Agent personalities",[11,854,855,856,858],{},"Every agent has a ",[33,857,181],{}," — a system prompt that controls how it thinks, responds, and behaves. A personality might say: \"You are a senior software engineer who reviews pull requests with a focus on security and clarity.\"",[11,860,861],{},"The personality is optional, but it makes a significant difference. An agent without a personality will be generic. One with a well-written personality will feel purpose-built for your workflow.",{"title":284,"searchDepth":285,"depth":285,"links":863},[864,865,866],{"id":742,"depth":285,"text":743},{"id":775,"depth":285,"text":776},{"id":851,"depth":285,"text":852},"agents","Agents are AI assistants that plan, act, and complete tasks on your behalf — here's what they can do and how they work.",{},"\u002Fagents\u002Fwhat-are-agents",[872,873,874],"agents\u002Fcreate-an-agent","agents\u002Fset-up-an-agent-profile","conversations\u002Freview-plans-and-approvals",{"title":731,"description":868},"what-are-agents","help\u002Fagents\u002F01.what-are-agents","fQhtPN58rCzs7yUj7Wctzo3MqDyxvNS6F340jCdn6C8","What are agents? Agents are AI assistants that plan, act, and complete tasks on your behalf — here's what they can do and how they work.",1778463886552]