Skip to main content

telnyx-voice-gather-java

Collect DTMF input and speech from callers using standard gather or AI-powered gather. Build interactive voice menus and AI voice assistants. This skill provides Java SDK examples.

معلومات المصدر

المستودع
team-telnyx/telnyx-toolkit
آخر نشاط في المصدر
١٠ فبراير ٢٠٢٦ في ٢٠:٤١
لغة SKILL.md المكتشفة
الإنجليزية
النجوم
٣
التفرعات
٢

خيارات التثبيت

يُحدَّد Prompt الذي يراجع المصدر أولًا بشكل افتراضي. يمكنك التبديل إلى أمر مباشر أو تنزيل نسخة محلية.

مراجعة ملفات المصدر

اقرأ SKILL.md وأي ملفات مرافقة يعرضها SkillsMP قبل أن تقرر التثبيت.

عرض SKILL.md

SKILL.md
تعليمات المصدر · معاينة للقراءة فقط
name
telnyx-voice-gather-java
description
Collect DTMF input and speech from callers using standard gather or AI-powered gather. Build interactive voice menus and AI voice assistants. This skill provides Java SDK examples.
metadata
{"author":"telnyx","product":"voice-gather","language":"java","generated_by":"telnyx-ext-skills-generator"}
<!-- Auto-generated from Telnyx OpenAPI specs. Do not edit. --> # Telnyx Voice Gather - Java ## Installation ```text // See https://github.com/team-telnyx/telnyx-java for Maven/Gradle setup ``` ## Setup ```java import com.telnyx.sdk.client.TelnyxClient; import com.telnyx.sdk.client.okhttp.TelnyxOkHttpClient; TelnyxClient client = TelnyxOkHttpClient.fromEnv(); ``` All examples below assume `client` is already initialized as shown above. ## Add messages to AI Assistant Add messages to the conversation started by an AI assistant on the call. `POST /calls/{call_control_id}/actions/ai_assistant_add_messages` ```java import com.telnyx.sdk.models.calls.actions.ActionAddAiAssistantMessagesParams; import com.telnyx.sdk.models.calls.actions.ActionAddAiAssistantMessagesResponse; ActionAddAiAssistantMessagesResponse response = client.calls().actions().addAiAssistantMessages("call_control_id"); ``` ## Start AI Assistant Start an AI assistant on the call. `POST /calls/{call_control_id}/actions/ai_assistant_start` ```java import com.telnyx.sdk.models.calls.actions.ActionStartAiAssistantParams; import com.telnyx.sdk.models.calls.actions.ActionStartAiAssistantResponse; ActionStartAiAssistantResponse response = client.calls().actions().startAiAssistant("call_control_id"); ``` ## Stop AI Assistant Stop an AI assistant on the call. `POST /calls/{call_control_id}/actions/ai_assistant_stop` ```java import com.telnyx.sdk.models.calls.actions.ActionStopAiAssistantParams; import com.telnyx.sdk.models.calls.actions.ActionStopAiAssistantResponse; ActionStopAiAssistantResponse response = client.calls().actions().stopAiAssistant("call_control_id"); ``` ## Gather stop Stop current gather. `POST /calls/{call_control_id}/actions/gather_stop` ```java import com.telnyx.sdk.models.calls.actions.ActionStopGatherParams; import com.telnyx.sdk.models.calls.actions.ActionStopGatherResponse; ActionStopGatherResponse response = client.calls().actions().stopGather("call_control_id"); ``` ## Gather using AI Gather parameters defined in the request payload using a voice assistant. `POST /calls/{call_control_id}/actions/gather_using_ai` — Required: `parameters` ```java import com.telnyx.sdk.core.JsonValue; import com.telnyx.sdk.models.calls.actions.ActionGatherUsingAiParams; import com.telnyx.sdk.models.calls.actions.ActionGatherUsingAiResponse; ActionGatherUsingAiParams params = ActionGatherUsingAiParams.builder() .callControlId("call_control_id") .parameters(ActionGatherUsingAiParams.Parameters.builder() .putAdditionalProperty("properties", JsonValue.from("bar")) .putAdditionalProperty("required", JsonValue.from("bar")) .putAdditionalProperty("type", JsonValue.from("bar")) .build()) .build(); ActionGatherUsingAiResponse response = client.calls().actions().gatherUsingAi(params); ``` ## Gather using audio Play an audio file on the call until the required DTMF signals are gathered to build interactive menus. `POST /calls/{call_control_id}/actions/gather_using_audio` ```java import com.telnyx.sdk.models.calls.actions.ActionGatherUsingAudioParams; import com.telnyx.sdk.models.calls.actions.ActionGatherUsingAudioResponse; ActionGatherUsingAudioResponse response = client.calls().actions().gatherUsingAudio("call_control_id"); ``` ## Gather using speak Convert text to speech and play it on the call until the required DTMF signals are gathered to build interactive menus. `POST /calls/{call_control_id}/actions/gather_using_speak` — Required: `voice`, `payload` ```java import com.telnyx.sdk.models.calls.actions.ActionGatherUsingSpeakParams; import com.telnyx.sdk.models.calls.actions.ActionGatherUsingSpeakResponse; ActionGatherUsingSpeakParams params = ActionGatherUsingSpeakParams.builder() .callControlId("call_control_id") .payload("say this on call") .voice("male") .build(); ActionGatherUsingSpeakResponse response = client.calls().actions().gatherUsingSpeak(params); ``` ## Gather Gather DTMF signals to build interactive menus. `POST /calls/{call_control_id}/actions/gather` ```java import com.telnyx.sdk.models.calls.actions.ActionGatherParams; import com.telnyx.sdk.models.calls.actions.ActionGatherResponse; ActionGatherResponse response = client.calls().actions().gather("call_control_id"); ``` --- ## Webhooks The following webhook events are sent to your configured webhook URL. All webhooks include `telnyx-timestamp` and `telnyx-signature-ed25519` headers for verification (Standard Webhooks compatible). | Event | Description | |-------|-------------| | `callGatherEnded` | Call Gather Ended | | `CallAIGatherEnded` | Call AI Gather Ended | | `CallAIGatherMessageHistoryUpdated` | Call AI Gather Message History Updated | | `CallAIGatherPartialResults` | Call AI Gather Partial Results | | `CallConversationEnded` | Call Conversation Ended | | `callPlaybackStarted` | Call Playback Started | | `callPlaybackEnded` | Call Playback Ended | | `callDtmfReceived` | Call Dtmf Received |
عرض على GitHub