深层链接调用 -
这些是可选调用,可让您在
自定义的意向受众群体,让用户可以通过深层链接访问您网站的特定部分
对话。用户可以说出类似于“Ok Google, talk to”的指令
<显示名称>以匹配这些类型的
意图(例如,“Ok Google, talk to Facts about Google to get company”(Ok Google,与 Facts about Google 通话)
事实”)
[null,null,["最后更新时间 (UTC):2025-07-26。"],[[["\u003cp\u003eActions Builder allows you to define how users initiate conversations with your Actions using an invocation model.\u003c/p\u003e\n"],["\u003cp\u003eYou can use a main invocation, deep link invocations, or built-in intents to enable users to trigger your Actions.\u003c/p\u003e\n"],["\u003cp\u003eWhen an Action is invoked, you can handle the event through webhooks, static prompts, or transitions to specific scenes within your conversation model.\u003c/p\u003e\n"],["\u003cp\u003eDeep link invocations allow users to jump to specific parts of your Action by using targeted phrases in their requests.\u003c/p\u003e\n"],["\u003cp\u003eImplicit invocation allows users to trigger your Action without needing to know your project's display name, relying on contextual cues and system intents.\u003c/p\u003e\n"]]],["Invocation models define how users start conversations with Actions. Key invocation types include: **Main invocation**, triggered by \"Ok Google, talk to [display name]\"; **Deep link invocations**, using custom phrases like \"talk to [display name] to [deep link phrase]\"; and **Built-in intents**, allowing invocation without the display name, such as \"Ok Google, play a game\". Upon invocation, a handler can trigger a webhook, respond with a prompt, or transition to a scene for further logic processing.\n"],null,["# Build invocation models\n\nActions Builder Actions SDK\n\nAn invocation model specifies how users invoke the Actions in your project.\nWhen users invoke an Action, they begin a conversation with it, which is\ndefined by your [conversation model](/assistant/conversational/build/conversation). You can define invocations\nin the following ways:\n\n- [**Main invocation**](/assistant/conversational/build/main-invocation) - This\n is the only required invocation for your Actions project. Users say a phrase\n similar to *\"Ok Google, talk to \\\u003cdisplay name\\\u003e\"* to match this intent,\n which you typically handle by greeting the user with a welcome message.\n\n- [**Deep link invocations**](/assistant/conversational/build/deep-link-invocations) -\n These are optional invocations that let you define training phrases in a\n custom intent, so users can deep link into a specific part of your\n conversation. Users can say a phrase similar to *\"Ok Google, talk to\n \\\u003cdisplay name\\\u003e to \\\u003cdeep link phrase\\\u003e\"* to match these types of\n intents (for example, *\"Ok Google, talk to Facts about Google to get company\n facts\"*)\n\n- [**Built-in intents**](/assistant/conversational/build/built-in-intents) -\n These are optional invocations that let you take advantage of system-defined\n invocation models. By declaring support for a built-in intent, users can\n discover and invoke your Actions project without having to specify your\n display name. For example, users can say *\"Ok Google, play a game\"*, and\n Google Assistant can present your Action as an option for invocation.\n\n**Figure 1.** An invocation model consists of all the intents that are eligible to start conversations with your Actions. **Note:** Only intents that are defined in your invocation model are eligible for matching during invocation. User intents in your [conversation model](#conversation_model) aren't eligible for invocation.\n\nCreate the main invocation\n--------------------------\n\nEvery Actions project has a default, main invocation that acts as the default\nentry point into your Actions project.The language model for this invocation\nis associated with your project's display name, so users just need to say a\nphrase similar to *\"Ok Google, talk to \\\u003cdisplay name\\\u003e\"* to use this\ninvocation. When users invoke your Actions in this manner, the Assistant\nruntime matches the `actions.intent.MAIN` system intent.\n\nTo create the main invocation:\n\nCreate deep link invocations\n----------------------------\n\nTo build deep link invocations, you create global user intents with training\nphrases that define the user input that you want to set for the deep link.\nUsers say your display name in combination with the deep link phrase to carry\nout a deep link invocation (for example, *\"Ok Google, talk to \\\u003cdisplay\nname\\\u003e to \\\u003cdeep link phrase\\\u003e\"*)\n| **Note:** Intents that are associated with deep link invocations are also valid for matching during a conversation.\n\nCreate implicit invocations\n---------------------------\n\n| **Note:** Currently, only the `actions.intent.PLAY_GAME` system intent is supported. You must register for this intent in order to be eligible for implicit invocation if you built a game.\n\nImplicit invocation lets users invoke your Action without using your project's\ndisplay name. To find matching Actions, Google uses signals such as users\ntelling Assistant to do something that's similar to an invocation phrase for one\nof your configured intents, or when the user is in a context where your Actions\nwould be appropriate.\n\nHandle invocation events\n------------------------\n\nWhen Assistant matches one of your global intents, it triggers an event that you\ncan process with a handler. The following sections describe what your handler\ncan carry out (you can choose one or any combination of the three).\n\n### Trigger a webhook\n\nYour intent handler can delegate to a web service by triggering a webhook event.\nTo trigger a webhook and have a web service handle the invocation:\n| **Note:** See the [webhooks](../webhooks) documentation for more information on what you can do in your fulfillment.\n\n### Prompts\n\nYour intent handler can respond with a static prompt. Typically, prompts for\ninvocation intents welcome the user, orient them on how to use the Action, or\nrespond with the information the user needs immediately (one-shot Actions).\n\nTo handle an invocation intent match with a prompt:\n| **Note:** See the [prompts](../prompts) documentation for more information on what you can do with prompts.\n\n### Transition to a scene\n\nYour intent handler can transition to a scene to process the matched invocation\nintent. This gives you the most flexibility to carry out logic. To transition to\na scene:\n| **Note:** See the [scenes](/assistant/conversational/scenes) and [conversation models](/assistant/conversational/build/conversation) documentation for more information on how to build scenes."]]