[{"data":1,"prerenderedAt":916},["ShallowReactive",2],{"help-category-\u002Fgetting-started\u002Fstart-your-first-conversation":3,"help-article-\u002Fgetting-started\u002Fstart-your-first-conversation":4,"related-articles-\u002Fgetting-started\u002Fstart-your-first-conversation":228},[],{"id":5,"title":6,"body":7,"category":210,"description":211,"draft":212,"extension":213,"meta":214,"navigation":215,"order":216,"path":217,"relatedArticles":218,"seo":222,"slug":223,"stem":224,"updatedAt":225,"__hash__":226,"excerpt":211,"searchText":227},"help\u002Fhelp\u002Fgetting-started\u002F11.start-your-first-conversation.md","Start your first conversation",{"type":8,"value":9,"toc":201},"minimark",[10,14,19,22,54,58,61,81,84,91,103,108,112,115,134,137,141,148,151,154,158],[11,12,13],"p",{},"You've installed MultiClaw and connected your workspace. Now it's time to talk to your first agent. Open the app, pick an agent, and type your first message.",[15,16,18],"h2",{"id":17},"before-you-start","Before you start",[11,20,21],{},"Make sure two things are in place:",[23,24,25,43],"ul",{},[26,27,28,32,33,36,37,42],"li",{},[29,30,31],"strong",{},"Active connection",": Check the ",[29,34,35],{},"gateway status indicator"," in the top-right corner. It should be green. If it shows red, ",[38,39,41],"a",{"href":40},"\u002Fgetting-started\u002Fconnect-to-your-workspace","connect to your workspace"," first.",[26,44,45,48,49,53],{},[29,46,47],{},"At least one agent",": If you haven't created one yet, see ",[38,50,52],{"href":51},"\u002Fagents\u002Fcreate-an-agent","Create an agent",".",[15,55,57],{"id":56},"start-a-conversation","Start a conversation",[11,59,60],{},"When no conversation is active, you'll see a welcome greeting — \"What can I help you build today?\" — with the composer centered below it.",[62,63,64,71,74],"ol",{},[26,65,66,67,70],{},"In the composer toolbar, click the ",[29,68,69],{},"agent name"," to open a searchable list, then select the agent you want to use.",[26,72,73],{},"Type your message in the composer.",[26,75,76,77,80],{},"Press ",[29,78,79],{},"Enter"," to send.",[11,82,83],{},"After you send your first message, the composer moves to the bottom of the screen. The agent's reply streams in real time, word by word. When the stream finishes, type a follow-up message to continue the conversation.",[85,86,88],"callout",{"type":87},"tip",[11,89,90],{},"Not sure what to type? Start with something specific: \"Create a Python script that lists all files in a folder\" or \"Draft an email introducing our new product.\" The more context you give, the better the response.",[85,92,94],{"type":93},"note",[11,95,96,97,53],{},"When you send a message, its content is processed by the AI model provider you have configured (such as OpenAI, Anthropic, or Google). Those providers handle this data under their own privacy policies. For details on how MultiClaw handles your data, see the ",[38,98,102],{"href":99,"rel":100},"https:\u002F\u002Fmulticlaw.io\u002Fprivacy",[101],"nofollow","MultiClaw Privacy Policy",[85,104,105],{"type":87},[11,106,107],{},"You can have multiple conversations with the same agent. Each conversation is a separate thread with its own history.",[15,109,111],{"id":110},"start-a-new-conversation","Start a new conversation",[11,113,114],{},"To open a fresh conversation at any time:",[23,116,117,124],{},[26,118,119,120,123],{},"Click the ",[29,121,122],{},"pencil icon"," in the top-left corner of the chat area, or",[26,125,76,126,129,130,133],{},[29,127,128],{},"Cmd+N"," (Mac) \u002F ",[29,131,132],{},"Ctrl+N"," (Windows\u002FLinux).",[11,135,136],{},"The agent selector locks after you send the first message. You'll see a lock icon next to the agent name. To use a different agent, start a new conversation.",[15,138,140],{"id":139},"when-the-agent-shows-a-plan","When the agent shows a plan",[11,142,143,144,147],{},"If your request requires action — using the browser, editing files, or sending a message — the agent drafts a ",[29,145,146],{},"plan"," before doing anything. The plan appears in the conversation as a list of steps the agent intends to take.",[11,149,150],{},"Review the steps, then approve the plan to let the agent proceed. The agent does not proceed until you approve.",[11,152,153],{},"For simple questions — like \"What does this error mean?\" — no plan appears. The agent answers directly in the conversation.",[15,155,157],{"id":156},"keyboard-shortcuts","Keyboard shortcuts",[159,160,161,174],"table",{},[162,163,164],"thead",{},[165,166,167,171],"tr",{},[168,169,170],"th",{},"Action",[168,172,173],{},"Shortcut",[175,176,177,185,193],"tbody",{},[165,178,179,183],{},[180,181,182],"td",{},"Send message",[180,184,79],{},[165,186,187,190],{},[180,188,189],{},"New line in message",[180,191,192],{},"Shift+Enter",[165,194,195,198],{},[180,196,197],{},"New conversation",[180,199,200],{},"Cmd+N (Mac) \u002F Ctrl+N (Windows\u002FLinux)",{"title":202,"searchDepth":203,"depth":203,"links":204},"",2,[205,206,207,208,209],{"id":17,"depth":203,"text":18},{"id":56,"depth":203,"text":57},{"id":110,"depth":203,"text":111},{"id":139,"depth":203,"text":140},{"id":156,"depth":203,"text":157},"getting-started","Open MultiClaw, select an agent, type your message, and start working — your first conversation takes less than a minute.",false,"md",{},true,11,"\u002Fgetting-started\u002Fstart-your-first-conversation",[219,220,221],"getting-started\u002Fexplore-the-interface","conversations\u002Fquick-chat-vs-tasks","agents\u002Fwhat-are-agents",{"title":6,"description":211},"start-your-first-conversation","help\u002Fgetting-started\u002F11.start-your-first-conversation",null,"kwNrpiIErxvXfrlmdv1XwgtGp7vMGJ6CWe_mLrTurwU","Start your first conversation Open MultiClaw, select an agent, type your message, and start working — your first conversation takes less than a minute.",[229,673,760],{"id":230,"title":231,"body":232,"category":210,"description":661,"draft":212,"extension":213,"meta":662,"navigation":215,"order":663,"path":664,"relatedArticles":665,"seo":668,"slug":669,"stem":670,"updatedAt":225,"__hash__":671,"excerpt":661,"searchText":672},"help\u002Fhelp\u002Fgetting-started\u002F10.explore-the-interface.md","Explore the interface",{"type":8,"value":233,"toc":649},[234,257,261,264,270,283,288,361,366,398,402,405,411,447,454,464,468,479,484,488,494,500,506,511,514,532,539,543,549,567,571,574,579,597,602,623,636,640],[11,235,236,237,240,241,244,245,248,249,252,253,256],{},"Use this article to learn where things live before you start work. MultiClaw is one window with four ideas to remember: a ",[29,238,239],{},"navigation sidebar"," on the left, a ",[29,242,243],{},"title bar"," across the top, a ",[29,246,247],{},"main content"," column that changes with the screen you pick, and (on ",[29,250,251],{},"Conversations",") the ",[29,254,255],{},"Chats sidebar"," plus chat surface.",[15,258,260],{"id":259},"the-navigation-sidebar","The navigation sidebar",[11,262,263],{},"The sidebar is how you move between major areas of the app. Icons link to each screen; labels identify what each icon opens.",[11,265,266,269],{},[29,267,268],{},"On a wide window",", the rail stays narrow and shows icons only. Hover over an icon (or focus it) to see its label in a tooltip.",[11,271,272,275,276,279,280,282],{},[29,273,274],{},"On a small window",", the same links live behind ",[29,277,278],{},"Open menu"," (hamburger). Click ",[29,281,278],{}," to open the full navigation sheet, then pick a destination.",[11,284,285],{},[29,286,287],{},"Top navigation:",[159,289,290,300],{},[162,291,292],{},[165,293,294,297],{},[168,295,296],{},"Section",[168,298,299],{},"What it does",[175,301,302,311,321,331,341,351],{},[165,303,304,308],{},[180,305,306],{},[29,307,251],{},[180,309,310],{},"Start and continue chats with your agents. This is your main working area.",[165,312,313,318],{},[180,314,315],{},[29,316,317],{},"My Agents",[180,319,320],{},"Create, configure, and manage your agents.",[165,322,323,328],{},[180,324,325],{},[29,326,327],{},"Instances",[180,329,330],{},"Start, stop, and monitor your cloud desktops.",[165,332,333,338],{},[180,334,335],{},[29,336,337],{},"Schedule",[180,339,340],{},"Schedule tasks to run automatically.",[165,342,343,348],{},[180,344,345],{},[29,346,347],{},"MCP Servers",[180,349,350],{},"Manage MCP servers that give agents access to external tools.",[165,352,353,358],{},[180,354,355],{},[29,356,357],{},"Skills",[180,359,360],{},"Browse, record, and manage reusable agent procedures.",[11,362,363],{},[29,364,365],{},"Bottom navigation:",[159,367,368,376],{},[162,369,370],{},[165,371,372,374],{},[168,373,296],{},[168,375,299],{},[175,377,378,388],{},[165,379,380,385],{},[180,381,382],{},[29,383,384],{},"Settings",[180,386,387],{},"Configure app preferences, appearance, gateway, and more.",[165,389,390,395],{},[180,391,392],{},[29,393,394],{},"Help",[180,396,397],{},"Open the help center.",[15,399,401],{"id":400},"the-title-bar","The title bar",[11,403,404],{},"The narrow strip at the very top shows edition and connection context. You can click and drag it to move the window.",[11,406,407,410],{},[29,408,409],{},"Companion (local OpenClaw) builds"," show, on the right when you are not connected to a cloud desktop:",[23,412,413,419,429],{},[26,414,415,418],{},[29,416,417],{},"Edition label"," so you know you are on the Companion build.",[26,420,421,424,425,53],{},[29,422,423],{},"Gateway port"," when the gateway is up, for example ",[426,427,428],"code",{},"Port 18789",[26,430,431,434,435,438,439,442,443,446],{},[29,432,433],{},"Gateway status"," with text ",[29,436,437],{},"Connected",", ",[29,440,441],{},"Offline",", or ",[29,444,445],{},"Reconnecting...",". A colored dot next to the label matches that state (green when connected, red when offline, amber while reconnecting). Open the status control whenever you need to start, stop, or restart the gateway.",[11,448,449,450,453],{},"When you ",[29,451,452],{},"are"," connected to a cloud desktop instance, that side of the bar shows the instance name and cloud controls instead of the local port and gateway row. Use the menu on the instance control to change instance or disconnect back to local work.",[11,455,456,459,460,463],{},[29,457,458],{},"Enterprise builds"," show ",[29,461,462],{},"MultiClaw Enterprise"," plus your signed-in organization and instance connection status on the right, tuned for cloud workspaces rather than the local gateway strip.",[15,465,467],{"id":466},"the-main-content-column","The main content column",[11,469,470,471,473,474,438,476,478],{},"Everything to the right of the navigation rail is the ",[29,472,247],{}," column. The navigation sidebar only switches routes: each item above replaces this column with that screen (",[29,475,317],{},[29,477,327],{},", and so on).",[11,480,481,483],{},[29,482,251],{}," fills the column with the chat layout described in the next section. Other items show their own pages (agent list, instance list, scheduler, and so on) in the same space.",[15,485,487],{"id":486},"the-chat-area-conversations","The chat area (Conversations)",[11,489,490,491,493],{},"When ",[29,492,251],{}," is active, the main column is your chat workspace. It behaves in two phases.",[11,495,496,499],{},[29,497,498],{},"Welcome phase"," — before the first message in a new flow, you see a centered greeting and the composer ready for input.",[11,501,502,505],{},[29,503,504],{},"Active phase"," — after messages exist, the transcript scrolls above the composer.",[507,508,510],"h3",{"id":509},"floating-controls","Floating controls",[11,512,513],{},"Two buttons stay in the top-left of the chat surface:",[23,515,516,524],{},[26,517,518,521,522,53],{},[29,519,520],{},"Sidebar toggle"," (panel icon): shows or hides the ",[29,523,255],{},[26,525,526,529,530,53],{},[29,527,528],{},"New Conversation"," (pencil icon): starts a new conversation without leaving ",[29,531,251],{},[11,533,534,535,538],{},"When you are connected to a cloud instance and a conversation is active, a control in the top-right opens the ",[29,536,537],{},"remote browser"," view for that instance.",[507,540,542],{"id":541},"the-chats-sidebar","The Chats sidebar",[11,544,545,546,548],{},"The ",[29,547,255],{}," slides beside the message area when you show it; it can start hidden so the transcript has more room. Inside you will find:",[23,550,551,558,564],{},[26,552,553,554,557],{},"A ",[29,555,556],{},"Chats"," heading.",[26,559,545,560,563],{},[29,561,562],{},"Search conversations..."," field, which filters the conversation list so you can find past chats.",[26,565,566],{},"Conversations ordered with the most recent first.",[507,568,570],{"id":569},"message-composer","Message composer",[11,572,573],{},"The composer sits at the bottom of the chat column.",[11,575,576],{},[29,577,578],{},"Left side of the toolbar:",[23,580,581,587],{},[26,582,583,586],{},[29,584,585],{},"Agent selector"," — choose which agent handles your messages. The selector can lock during an active conversation (lock icon) so you do not switch agents mid-session.",[26,588,589,592,593,596],{},[29,590,591],{},"Model selector"," — when you are working ",[29,594,595],{},"locally"," (not through a connected cloud desktop), choose the language model here. When you are connected to an instance, model choice follows the cloud agent setup, so this control is not part of the cloud composer.",[11,598,599],{},[29,600,601],{},"Right side of the toolbar:",[23,603,604,610],{},[26,605,606,609],{},[29,607,608],{},"Attach file"," — add files from disk; previews appear above the text field when something is attached.",[26,611,612,615,616,619,620,622],{},[29,613,614],{},"Send"," — sends the message. While the agent is generating a reply, ",[29,617,618],{},"Stop"," replaces ",[29,621,614],{}," so you can cancel generation.",[11,624,625,626,629,630,632,633,635],{},"Type in the main text field. ",[29,627,628],{},"Shift"," + ",[29,631,79],{}," adds a line; ",[29,634,79],{}," alone sends (when sending is allowed).",[507,637,639],{"id":638},"when-you-are-done-skimming","When you are done skimming",[11,641,642,643,645,646,648],{},"You will recognize the layout: icons at the left edge, status and edition along the top drag strip, and ",[29,644,251],{}," showing either the welcome centering or transcript plus composer. Open the ",[29,647,255],{}," when you need history or search.",{"title":202,"searchDepth":203,"depth":203,"links":650},[651,652,653,654],{"id":259,"depth":203,"text":260},{"id":400,"depth":203,"text":401},{"id":466,"depth":203,"text":467},{"id":486,"depth":203,"text":487,"children":655},[656,658,659,660],{"id":509,"depth":657,"text":510},3,{"id":541,"depth":657,"text":542},{"id":569,"depth":657,"text":570},{"id":638,"depth":657,"text":639},"A mental map of the MultiClaw window: navigation sidebar, title bar, main content, conversations, Chats sidebar, and composer.",{},10,"\u002Fgetting-started\u002Fexplore-the-interface",[666,221,667],"getting-started\u002Fstart-your-first-conversation","skills-workflows\u002Fwhat-are-skills",{"title":231,"description":661},"explore-the-interface","help\u002Fgetting-started\u002F10.explore-the-interface","1gPlFnl7pWQtPbcch7lyqtNFv40-WUJiYOUSHvYl9YM","Explore the interface A mental map of the MultiClaw window: navigation sidebar, title bar, main content, conversations, Chats sidebar, and composer.",{"id":674,"title":675,"body":676,"category":749,"description":750,"draft":212,"extension":213,"meta":751,"navigation":215,"order":203,"path":752,"relatedArticles":753,"seo":756,"slug":683,"stem":757,"updatedAt":225,"__hash__":758,"excerpt":750,"searchText":759},"help\u002Fhelp\u002Fconversations\u002F02.quick-chat-vs-tasks.md","Quick chat vs tasks",{"type":8,"value":677,"toc":745},[678,681,684,726,733,737,740],[11,679,680],{},"When you send a message, the agent responds in one of two ways: it either replies conversationally — answering questions, brainstorming, summarising — or it begins taking action on your behalf. The difference comes down to whether you asked a question or asked the agent to act.",[15,682,675],{"id":683},"quick-chat-vs-tasks",[159,685,686,698],{},[162,687,688],{},[165,689,690,692,695],{},[168,691],{},[168,693,694],{},"Quick chat",[168,696,697],{},"Tasks",[175,699,700,713],{},[165,701,702,707,710],{},[180,703,704],{},[29,705,706],{},"Purpose",[180,708,709],{},"Questions, brainstorming, summaries, discussion",[180,711,712],{},"Taking action — browsing, file management, sending messages, running a workflow",[165,714,715,720,723],{},[180,716,717],{},[29,718,719],{},"When to use",[180,721,722],{},"You want an answer or want to think something through",[180,724,725],{},"You want the agent to take action on your behalf",[11,727,728],{},[729,730],"img",{"alt":731,"src":732},"Quick chat exchange — the agent provides a conversational answer to a question","\u002Fimages\u002Fdemos\u002Fquick-chat-reply.gif",[15,734,736],{"id":735},"how-the-agent-decides","How the agent decides",[11,738,739],{},"The agent reads your message and determines whether you want a conversational reply or action. A question like \"What's a good subject line for this email?\" gets a direct reply. A request like \"Find all invoices from last quarter\" will have the agent start working through the task.",[85,741,742],{"type":87},[11,743,744],{},"If you want action, make sure your message clearly describes what you want done. If the agent responds conversationally to an action request, rephrase your message more explicitly.",{"title":202,"searchDepth":203,"depth":203,"links":746},[747,748],{"id":683,"depth":203,"text":675},{"id":735,"depth":203,"text":736},"conversations","Understand when the agent replies conversationally and when it takes action on your behalf.",{},"\u002Fconversations\u002Fquick-chat-vs-tasks",[754,755,221],"conversations\u002Fstart-a-conversation","conversations\u002Fconversation-history",{"title":675,"description":750},"help\u002Fconversations\u002F02.quick-chat-vs-tasks","vLX2DOxFmzgI7nSgS2v5cbu1_frzRCN5H3SLkPCziaA","Quick chat vs tasks Understand when the agent replies conversationally and when it takes action on your behalf.",{"id":761,"title":762,"body":763,"category":902,"description":903,"draft":212,"extension":213,"meta":904,"navigation":215,"order":905,"path":906,"relatedArticles":907,"seo":911,"slug":912,"stem":913,"updatedAt":225,"__hash__":914,"excerpt":903,"searchText":915},"help\u002Fhelp\u002Fagents\u002F01.what-are-agents.md","What are agents?",{"type":8,"value":764,"toc":897},[765,768,771,775,799,804,808,880,883,887,894],[11,766,767],{},"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.",[11,769,770],{},"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.",[15,772,774],{"id":773},"what-agents-can-do","What agents can do",[23,776,777,780,783,786,789,792],{},[26,778,779],{},"Answer questions and hold conversations",[26,781,782],{},"Draft step-by-step plans for tasks and wait for your approval",[26,784,785],{},"Execute tasks on a cloud desktop — browsing the web, filling forms, editing files",[26,787,788],{},"Use MCP tools to connect to apps like GitHub, Slack, Google Drive, and others",[26,790,791],{},"Run recorded browser workflows and reusable skills — each run is subject to your approval before execution",[26,793,794,795,798],{},"Delegate subtasks to other agents (",[29,796,797],{},"subagents",") — delegation is included in the plan you approve before execution begins",[85,800,801],{"type":93},[11,802,803],{},"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.",[15,805,807],{"id":806},"local-agents-vs-cloud-agents","Local agents vs cloud agents",[159,809,810,822],{},[162,811,812],{},[165,813,814,816,819],{},[168,815],{},[168,817,818],{},"Local agent",[168,820,821],{},"Cloud agent",[175,823,824,841,854,867],{},[165,825,826,831,838],{},[180,827,828],{},[29,829,830],{},"Where stored",[180,832,833,834,837],{},"Your machine (",[426,835,836],{},"~\u002F.openclaw\u002F",")",[180,839,840],{},"MultiClaw Cloud",[165,842,843,848,851],{},[180,844,845],{},[29,846,847],{},"Who can access it",[180,849,850],{},"Only you, on that device",[180,852,853],{},"Any workspace member with permission",[165,855,856,861,864],{},[180,857,858],{},[29,859,860],{},"Syncs across devices",[180,862,863],{},"No",[180,865,866],{},"Yes",[165,868,869,874,877],{},[180,870,871],{},[29,872,873],{},"Runs 24\u002F7",[180,875,876],{},"Only while your machine is on",[180,878,879],{},"Yes, on a cloud desktop",[11,881,882],{},"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.",[15,884,886],{"id":885},"agent-personalities","Agent personalities",[11,888,889,890,893],{},"Every agent has a ",[29,891,892],{},"personality"," — 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.\"",[11,895,896],{},"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":202,"searchDepth":203,"depth":203,"links":898},[899,900,901],{"id":773,"depth":203,"text":774},{"id":806,"depth":203,"text":807},{"id":885,"depth":203,"text":886},"agents","Agents are AI assistants that plan, act, and complete tasks on your behalf — here's what they can do and how they work.",{},1,"\u002Fagents\u002Fwhat-are-agents",[908,909,910],"agents\u002Fcreate-an-agent","agents\u002Fset-up-an-agent-profile","conversations\u002Freview-plans-and-approvals",{"title":762,"description":903},"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.",1778463886548]