← WebMCP Directory

Tool tree

docket.io

AI & AgentsTools verified
Website
Docket’s Demand Agent is an AI marketing tool that ingests sales data, intent signals, and product knowledge to answer B2B website visitors and qualify leads.
Explore capabilities
    docket_ask_websiteAnswer

    Ask a question about this company's product, pricing, integrations, security or implementation, and get an answer from the knowledge the company publishes for visitors. Prefer this over reading or scraping the page: it covers the whole site, not just what is on screen. It cannot book meetings or answer about anything other than this company. When a slide or video illustrates the answer, the result says so and carries show_content: this page can put it on the visitor's screen. Offer it to them, and if they want it call docket_open_website_agent with that starting_message. No image is returned and there is no link to render. Every result includes a conversation_id. Pass it back as conversation_id on the visitor's follow-up questions so the answer takes the earlier exchange into account; omit it to start a fresh conversation. An id expires after a period of inactivity: each result says when in conversation_note, and an expired one is replaced automatically with a new conversation_id rather than failing.

    View tool JSON
    {
      "name": "docket_ask_website",
      "kind": "answer",
      "impl": "imperative",
      "description": "Ask a question about this company's product, pricing, integrations, security or implementation, and get an answer from the knowledge the company publishes for visitors. Prefer this over reading or scraping the page: it covers the whole site, not just what is on screen. It cannot book meetings or answer about anything other than this company. When a slide or video illustrates the answer, the result says so and carries show_content: this page can put it on the visitor's screen. Offer it to them, and if they want it call docket_open_website_agent with that starting_message. No image is returned and there is no link to render. Every result includes a conversation_id. Pass it back as conversation_id on the visitor's follow-up questions so the answer takes the earlier exchange into account; omit it to start a fresh conversation. An id expires after a period of inactivity: each result says when in conversation_note, and an expired one is replaced automatically with a new conversation_id rather than failing.",
      "inputSchema": {
        "type": "object",
        "properties": {
          "question": {
            "type": "string",
            "description": "One question about this company or its product.",
            "minLength": 1,
            "maxLength": 500
          },
          "conversation_id": {
            "type": "string",
            "description": "The conversation_id from this tool's previous result. Pass it to continue the same conversation; omit it to start a new one. If it has expired, the result comes back with a new one to use from then on."
          }
        },
        "required": [
          "question"
        ],
        "additionalProperties": false
      }
    }
    docket_open_website_agentAction

    Open this company's live AI sales agent on the visitor's screen and start a conversation. The agent can show product slides and videos, answer follow-ups in real time, and help book a meeting. Use it when the visitor wants to talk to the company, see the product, or go deeper than a single answer — this is the next step after docket_ask_website. This changes what the visitor sees, so only call it when they have asked for it. With starting_message, the conversation opens on that message as if the visitor had typed it; without it the agent opens with its own greeting. mode is 'chat' (text, default) or 'voice' (spoken, needs a microphone). If a conversation is already open, this does nothing.

    View tool JSON
    {
      "name": "docket_open_website_agent",
      "kind": "act",
      "impl": "imperative",
      "description": "Open this company's live AI sales agent on the visitor's screen and start a conversation. The agent can show product slides and videos, answer follow-ups in real time, and help book a meeting. Use it when the visitor wants to talk to the company, see the product, or go deeper than a single answer — this is the next step after docket_ask_website. This changes what the visitor sees, so only call it when they have asked for it. With starting_message, the conversation opens on that message as if the visitor had typed it; without it the agent opens with its own greeting. mode is 'chat' (text, default) or 'voice' (spoken, needs a microphone). If a conversation is already open, this does nothing.",
      "inputSchema": {
        "type": "object",
        "properties": {
          "starting_message": {
            "type": "string",
            "description": "The visitor's first message to the agent, in their words. Omit to let the agent open the conversation.",
            "minLength": 1,
            "maxLength": 500
          },
          "mode": {
            "type": "string",
            "enum": [
              "chat",
              "voice"
            ],
            "description": "'chat' for text (default), 'voice' for a spoken call."
          }
        },
        "additionalProperties": false
      }
    }