# 应用(bot) API > 全部文档的索引见 [llms.txt](https://ark.volcengine.com/llms.txt)；在任一文档页面地址后追加 `.md` 可获取该页的 Markdown 版本。 ` POST https://ark.cn-beijing.volces.com/api/v3/bots/chat/completions` [运行](https://api.volcengine.com/api-explorer/?action=BotsChatCompletions&data=%7B%7D&groupName=%E5%BA%94%E7%94%A8Bot%20API&query=%7B%7D&serviceCode=ark&version=2024-01-01) 本文介绍方舟应用 API 的输入输出参数，供您使用接口时查阅字段含义。 **选项：快速入口** ![图片](https://portal.volccdn.com/obj/volcfe/cloud-universal-doc/upload_b9c82890e851fc10cc31f48f9065abc6.png) [体验中心](https://ark.volcengine.com/region:cn-beijing/application) ![图片](https://portal.volccdn.com/obj/volcfe/cloud-universal-doc/upload_2abecd05ca2779567c6d32f0ddc7874d.png) [模型列表](/docs/model-list#文本生成能力) ![图片](https://portal.volccdn.com/obj/volcfe/cloud-universal-doc/upload_a5fdd3028d35cc512a10bd71b982b6eb.png) [模型计费](/docs/model-pricing#%E5%A4%A7%E8%AF%AD%E8%A8%80%E6%A8%A1%E5%9E%8B) ![图片](https://portal.volccdn.com/obj/volcfe/cloud-universal-doc/upload_afbcf38bdec05c05089d5de5c3fd8fc8.png) [API Key](https://ark.volcengine.com/region:cn-beijing/apiKey?apikey=%7B%7D) ![图片](https://portal.volccdn.com/obj/volcfe/cloud-universal-doc/upload_57d0bca8e0d122ab1191b40101b5df75.png) [获取Bot ID](/docs/no-code-app-guide) ![图片](https://portal.volccdn.com/obj/volcfe/cloud-universal-doc/upload_f45b5cd5863d1eed3bc3c81b9af54407.png) [接口文档](/docs/bot-api) ![图片](https://portal.volccdn.com/obj/volcfe/cloud-universal-doc/upload_1609c71a747f84df24be1e6421ce58f0.png) [常见问题](/docs/modelark-faq) ![图片](https://portal.volccdn.com/obj/volcfe/cloud-universal-doc/upload_bef4bc3de3535ee19d0c5d6c37b0ffdd.png) [开通模型](https://ark.volcengine.com/region:cn-beijing/openManagement?LLM=%7B%7D&OpenTokenDrawer=false) **选项：鉴权说明** 本接口支持 API Key 鉴权，详见 [鉴权认证方式](/docs/base-url-and-authentication)。 > 如需使用 Access Key 来鉴权，推荐使用 SDK 的方式，具体请参见 [SDK概述](/docs/install-and-upgrade-sdk)。 --- ## 请求参数 > 跳转 [响应参数](/docs/bot-api#Qu59cel0) ### 请求体 * **model** `string` `必选` 您需要调用的应用的 ID （Bot ID），[获取 Bot ID](/docs/no-code-app-guide) 。 * **messages** `object[]` `必选` 到目前为止的对话组成的消息列表。不同模型支持不同类型的消息，如文本、图片等。 **选项：System message** `object` 系统消息，开发人员提供的指令，模型应遵循这些指令。如模型扮演的角色或者目标等。 * messages.**role** `string` `必选` 发送消息的角色，此处应为`system`。 * messages.**content** `string / object[]` `必选` 系统信息内容。 **选项：纯文本消息内容** \`string\` 纯文本的消息内容。 **选项：多模态消息内容** \`object[]\` 视觉理解模型等多模态模型、部分大语言模型支持此类型。 **文本消息部分** \`object\` 多模态消息中，内容文本输入。[视觉理解模型](/docs/model-list#视觉理解)、部分大语言模型支持此类型消息。 * messages.content.**text **\`string\` \`必选\` 文本消息内容部分。 * messages.content.**type **\`string\` \`必选\` 文本消息类型，此次应为 \`text\`。 **图像消息部分** \`object\` 多模态消息中，图像内容部分。[视觉理解模型](/docs/model-list#视觉理解)支持此类型消息。 * messages.content.**image_url **\`object\` \`必选\` 图片消息的内容部分。 * messages.content.image_url.**url **\`string\` \`必选\` 支持传入图片链接或图片的Base64编码，不同模型支持图片大小略有不同，具体请参见[使用说明](/docs/image-understanding#7a123cd1)。 * 传入图片URL：传入图片的可访问链接，推荐使用 TOS（火山引擎对象存储）存储图片，并生成图片链接。 * 传入Base64编码：请遵循格式\`data:image/<图片格式>;base64,\`，可见[示例](/docs/image-understanding#477e51ce)。 * messages.content.image_url.**detail **\`string\` \`默认值 auto\` 支持手动设置图片的质量，取值范围\`high\`、\`low\`、\`auto\`。 * \`high\`：高细节模式，适用于需要理解图像细节信息的场景，如对图像的多个局部信息/特征提取、复杂/丰富细节的图像理解等场景，理解更全面。 * \`low\`：低细节模式，适用于简单的图像分类/识别、整体内容理解/描述等场景，理解更快速。 * \`auto\`：默认模式，不同模型选择的模式略有不同，具体请参见[理解图像的深度控制](/docs/image-understanding#bf4d9224)。 * messages.content.**type **\`string\` \`必选\` 图像消息类型，此次应为 \`image_url\`。 **选项：User message** `object` 用户发送的消息，包含提示或附加上下文信息。 * messages.**role** `string` `必选` 发送消息的角色，此处应为`user`。 * messages.**content** `string / object[]` `必选` 用户信息内容。 **选项：纯文本消费内容** \`string\` 大语言模型支持此类型。 **选项：多模态消息内容** \`object[]\` 视觉理解模型等多模态模型、部分大语言模型支持此类型。 **文本消息部分** \`object\` 多模态消息中，内容文本输入。[视觉理解模型](/docs/model-list#视觉理解)、部分大语言模型支持此类型消息。 * messages.content.**text **\`string\` \`必选\` 文本消息内容部分。 * messages.content.**type **\`string\` \`必选\` 文本消息类型，此次应为 \`text\`。 **图像消息部分** \`object\` 多模态消息中，图像内容部分。[视觉理解模型](/docs/model-list#视觉理解)支持此类型消息。 * messages.content.**image_url **\`object\` \`必选\` 图片消息的内容部分。 * messages.content.image_url.**url **\`string\` \`必选\` 支持传入图片链接或图片的Base64编码，不同模型支持图片大小略有不同，具体请参见[使用说明](/docs/image-understanding#7a123cd1)。 * 传入图片URL：传入图片的可访问链接，推荐使用 TOS（火山引擎对象存储）存储图片，并生成图片链接。 * 传入Base64编码：请遵循格式\`data:image/<图片格式>;base64,\`，可见[示例](/docs/image-understanding#477e51ce)。 * messages.content.image_url.**detail **\`string / null\` \`默认值 auto\` 支持手动设置图片的质量，取值范围\`high\`、\`low\`、\`auto\`。 * \`high\`：高细节模式，适用于需要理解图像细节信息的场景，如对图像的多个局部信息/特征提取、复杂/丰富细节的图像理解等场景，理解更全面。 * \`low\`：低细节模式，适用于简单的图像分类/识别、整体内容理解/描述等场景，理解更快速。 * \`auto\`：默认模式，不同模型选择的模式略有不同，具体请参见[理解图像的深度控制](/docs/image-understanding#bf4d9224)。 * messages.content.**type **\`string\` \`必选\` 图像消息类型，此次应为 \`image_url\`。 * messages.**name** `string` 发送此消息的角色的姓名。用于区别同一个角色但是不同主体发送的消息。 **选项：Assistant message** `object` 模型响应用户消息而回复的消息。 > **说明** > > messages.**content** 与 messages.**tool_calls** 字段二者至少填写其一。 * messages.**role** `string` `必选` 发送消息的角色，此处应为`assistant`。 * messages.**content** `string / array` 模型回复的消息。 * messages.**tool_calls** `object[]` 模型回复的工具调用信息。 * messages.tool_calls**.function **`object` `必选` 模型调用工具对应的函数信息。 * messages.tool_calls **.** function.**name **`string` `必选` 模型需要调用的函数名称。 * messages.tool_calls **.** function.**arguments **`string` `必选` 模型生成的用于调用函数的参数，JSON 格式。 > **说明** > > 模型并不总是生成有效的 JSON，并且可能会虚构出一些您的函数参数规范中未定义的参数。在调用函数之前，请在您的代码中验证这些参数是否有效。 * messages.tool_calls**.id **`string` `必选` 调用的工具的 ID。 * messages.tool_calls**.type **`string` `必选` 工具类型，当前仅支持`function`。 **选项：Tool message** `object` 模型调用工具的消息。 * messages.**role** `string` `必选` 发送消息的角色，此处应为`tool`。 * messages.**content** `string / array` `必选` 工具返回的消息。 * messages.**tool_call_id **`string` `必选` 模型调用的工具的 ID。 * **thinking** `object` `默认值 {"type":"enabled"}` 控制模型是否开启深度思考模式。默认开启深度思考模式，可以手动关闭。 > 支持此字段的模型以及使用示例请参见 [文档](/docs/deep-thinking#fa3f44fa)。属性 --- thinking.**type** `string` `必选` 取值范围：`enabled`， `disabled`，`auto`。 * `enabled`：开启思考模式，模型一定先思考后回答。 * `disabled`：关闭思考模式，模型直接回答问题，不会进行思考。 * `auto`：自动思考模式，模型根据问题自主判断是否需要思考，简单题目直接回答。 * **stream** `boolean / null` `默认值 false` 响应内容是否流式返回： * `false`：模型生成完所有内容后一次性返回结果。 * `true`：按 SSE 协议逐块返回模型生成内容，并以一条 `data: [DONE] `消息结束。 * **stream_options** `object / null` `默认值 null` 流式响应的选项。当 **stream** 为 `true` 时可设置 **stream_options** 字段。 * stream_options.**include_usage** `boolean / null` `默认值 false` 是否包含本次请求的 token 用量统计信息。 * `false`：不返回 token 用量信息。 * `true`：在 `data: [DONE]` 消息之前会返回一个额外的块，此块上的 **usage** 字段代表整个请求的 token 用量，**choices** 字段为空数组。所有其他块也将包含 **usage** 字段，但值为 `null`。 * **max_tokens** `integer / null` `默认值 4096` 模型回复最大长度（单位 token），取值范围各个模型不同，详细见[模型列表](/docs/model-list)。输入 token 和输出 token 的总长度还受模型的上下文长度限制。 * **stop** `string / string[] / null` `默认值 null` 模型遇到 stop 字段所指定的字符串时将停止继续生成，这个词语本身不会输出。 * 默认最多支持 `4` 个停止词。 * `deepseek-v4-pro-ga-260813`、`deepseek-v4-flash-ga-260731` 最多支持 `16` 个停止词。 > [深度思考能力模型](/docs/model-list#76df97c8) 不支持该字段；上述 DeepSeek V4 模型除外。示例：`["你好", "天气"]` * **frequency_penalty** `float / null` `默认值 0` 取值范围为 [-2.0, 2.0]。频率惩罚系数。如果值为正，会根据新 token 在文本中的出现频率对其进行惩罚，从而降低模型逐字重复的可能性。 * **presence_penalty** `float / null` `默认值 0` 取值范围为 [-2.0, 2.0]。存在惩罚系数。如果值为正，会根据新 token 到目前为止是否出现在文本中对其进行惩罚，从而增加模型谈论新主题的可能性。 * **temperature** `float / null` `默认值 1` 取值范围为 [0, 2]。采样温度。控制了生成文本时对每个候选词的概率分布进行平滑的程度。。当取值为 0 时模型仅考虑对数概率最大的一个 token。较高的值（如 0.8）会使输出更加随机，而较低的值（如 0.2）会使输出更加集中确定。通常建议仅调整 temperature 或 top_p 其中之一，不建议两者都修改。 * **top_p** `float / null` `默认值 0.7` 取值范围为 [0, 1]。核采样概率阈值。模型会考虑概率质量在 top_p 内的 token 结果。当取值为 0 时模型仅考虑对数概率最大的一个 token。 0.1 意味着只考虑概率质量最高的前 10% 的 token，取值越大生成的随机性越高，取值越低生成的确定性越高。通常建议仅调整 temperature 或 top_p 其中之一，不建议两者都修改。 * **logprobs** `boolean / null` `默认值 false` 是否返回输出 tokens 的对数概率。 * `false`：不返回对数概率信息。 * `true`：返回消息内容中每个输出 token 的对数概率。 * **top_logprobs** `integer / null` `默认值 0` 取值范围为 [0, 20]。指定每个输出 token 位置最有可能返回的 token 数量，每个 token 都有关联的对数概率。仅当 **logprobs** 为`true` 时可以设置** top_logprobs** 参数。 * **logit_bias** `map / null` `默认值 null` 调整指定 token 在模型输出内容中出现的概率，使模型生成的内容更加符合特定的偏好。**logit_bias** 字段接受一个 map 值，其中每个键为词表中的 token ID（使用 tokenization 接口获取），每个值为该 token 的偏差值，取值范围为 [-100, 100]。 -1 会减少选择的可能性，1 会增加选择的可能性；-100 会完全禁止选择该 token，100 会导致仅可选择该 token。该参数的实际效果可能因模型而异。 `{ "1234": -100}` * **tools** `object[] / null` `默认值 null` 模型可以调用的工具列表。目前仅函数作为工具被支持。用这个来提供模型可能为其生成 JSON 输入的函数列表。支持该字段的模型请参见 [文档](/docs/model-list#f44ceef7)。 * tools.**type** `string` `必选` 工具类型，当前仅支持 `function`。 * tools.**function** `object` `必选` 模型可以调用的类型为`function`的工具列表。 * tools.function.**name** `string` `必选` 调用的函数的名称。 * tools.function.**description** `string` `必选` 调用的函数的描述，大模型会使用它来判断是否调用这个函数。 * tools.function.**parameters** `object` `必选` 函数请求参数，以 JSON Schema 格式描述。具体格式请参考 [JSON Schema](https://json-schema.org/understanding-json-schema) 文档。 ```json { "type": "object", "properties": { "location": { "type": "string", "description": "城市，如：北京" } }, "required": ["location"] } ``` * **metadata** `object / null` `默认值 null` 额外参数。 * metadata.**group_chat_config** `object` `必选` 运行时动态传入的角色配置。 * metadata.group_chat_config.**characters** `object[]` `必选` 群聊角色列表。 * metadata.group_chat_config.characters.**name** `string` `必选` 群聊角色名称。 * metadata.group_chat_config.characters.**system_prompt** `string` `必选` 群聊角色设定，用于告知应用需要扮演的角色。 * metadata.group_chat_config.characters.**model_desc** `object` `必选` 模型说明。 * metadata.group_chat_config.characters.model_desc.**endpoint_id** `string` `必选` 您创建的 [推理接入点](https://ark.volcengine.com/region:cn-beijing/docs/create-standard-inference-endpoint) ID。 * metadata.group_chat_config.**description** `string / null` `默认值 null` 群聊场景描述，可设定群聊的主题、时间地点、事件场景、用户扮演的角色等信息。 * metadata.group_chat_config.**user_name** `string / null` `默认值用户` 表示“我”所扮演的角色名称，默认值为“用户”。 * metadata.**user_info** `string / null` `默认值 null` 联网的智能体可使用，如查询天气场景，前端可以直接通过该字段传地点信息给模型，方便查对应地点的天气。传入的信息需满足条件：可以被反序列化成 json 的字符串，并需包含 `city` 和 `district` 两个字段。 * metadata.**emit_intention_signal_extra** `string / null` `默认值 "false"` 值是 `"true"` 时，会中途返回intention状态 "正在搜索"。 * metadata.**target_character_name** `string / null` `默认值 null` 群聊Bot对话时填写，指定本次要发言的角色名，必须是存在于`characters`里的角色。 ## 响应参数 > 跳转 [请求参数](/docs/bot-api#RxN8G2nH) **选项：非流式调用返回** * **id** `string` 本次请求的唯一标识。 * **model** `string` 本次请求实际使用的模型名称和版本。 doubao 1.5 代模型的模型名称格式为 `doubao-1-5-**`。如调用部署doubao-1.5-pro-32k 250115模型的推理接入点，返回model字段信息doubao-1-5-pro-32k-250115。 * **created** `integer` 本次请求创建时间的 Unix 时间戳（秒）。 * **object** `string` 固定为 `chat.completion`。 * **choices** `object[]` 本次请求的模型输出内容。 * choices.**index **`integer` 当前元素在 **choices** 列表的索引。 * choices.**finish_reason **`string` 模型停止生成 token 的原因。取值范围： * `stop`：模型输出自然结束，或因命中请求参数 stop 中指定的字段而被截断。 * `length`：模型输出因达到模型输出限制而被截断，有下面几种原因： * 触发`max_token`限制（回答内容的长度限制）。 * 触发`max_completion_tokens`限制（思维链内容+回答内容的长度限制）。 * 触发`context_window` 限制（输入内容+思维链内容+回答内容的长度限制）。 * `content_filter`：模型输出被内容审核拦截。 * `tool_calls`：模型调用了工具。 * choices.**message **`object` 模型输出的内容。 * choices.message.**role **`string` 内容输出的角色，此处固定为 `assistant`。 * choices.message.**content **`string` 模型生成的消息内容。 * choices.message.**reasoning_content **`string / null` 模型处理问题的思维链内容。仅深度推理模型支持返回此字段，深度推理模型请参见 [支持模型](/docs/deep-thinking#fee31637)。 * choices.message.**tool_calls **`object[] / null` 模型生成的工具调用。 * choices.message.tool_calls.**i****d **`string` 调用的工具的 ID。 * choices.message.tool_calls.**type **`string` 工具类型，当前仅支持`function`。 * choices.message.tool_calls.**function **`object` 模型调用的函数。 * choices.message.tool_calls.function.**name **`string` 模型调用的函数的名称。 * choices.message.tool_calls.function.**arguments **`string` 模型生成的用于调用函数的参数，JSON 格式。模型并不总是生成有效的 JSON，并且可能会虚构出一些您的函数参数规范中未定义的参数。在调用函数之前，请在您的代码中验证这些参数是否有效。 * choices.**logprobs **`object / null` 当前内容的对数概率信息。 * choices.logprobs.**content **`object[] / null` message列表中每个 content 元素中的 token 对数概率信息。 * choices.logprobs.content.**token **`string` 当前 token。 * choices.logprobs.content.**bytes **`integer[] / null` 当前 token 的 UTF-8 值，格式为整数列表。当一个字符由多个 token 组成（表情符号或特殊字符等）时可以用于字符的编码和解码。如果 token 没有 UTF-8 值则为空。 * choices.logprobs.content.**logprob **`float` 当前 token 的对数概率。 * choices.logprobs.content.**top_logprobs **`object[]` 在当前 token 位置最有可能的标记及其对数概率的列表。在一些情况下，返回的数量可能比请求参数 top_logprobs 指定的数量要少。 * choices.logprobs.content.top_logprobs.**token **`string` 当前 token。 * choices.logprobs.content.top_logprobs.**bytes **`integer[] / null` 当前 token 的 UTF-8 值，格式为整数列表。当一个字符由多个 token 组成（表情符号或特殊字符等）时可以用于字符的编码和解码。如果 token 没有 UTF-8 值则为空。 * choices.logprobs.content.top_logprobs.**logprob **`float` 当前 token 的对数概率。 * choices.**moderation_hit_type **`string / null` 模型输出文字含有敏感信息时，会返回模型输出文字命中的风险分类标签。返回值及含义： * `severe_violation`：模型输出文字涉及严重违规。 * `violence`：模型输出文字涉及激进行为。注意：当前只有 [视觉理解模型](/docs/image-understanding#18cf565a) 支持返回该字段，且只有在方舟控制台 [接入点配置页面](https://ark.volcengine.com/region:cn-beijing/endpoint/create?customModelId=) 或者 [CreateEndpoint](/docs/create-endpoint-api) 接口中，将内容护栏方案（ModerationStrategy）设置为基础方案（Basic）时，才会返回风险分类标签。 * **bot_usage** `object` 本次请求的 token 用量。 * bot_usage.**model_usage **`object[]` 本次请求不同 endpoint 的 token 消耗。 * bot_usage.model_usage **.name** `string` 调用的模型 ID。 * bot_usage.model_usage **.prompt_tokens** `integer` 输入的 prompt token 数量，包含用户输入的提示词和调用插件返回的信息。 * bot_usage.model_usage **.completion_tokens** `integer` 模型生成的 token 数量。 * bot_usage.model_usage **.total_tokens** `integer` 本次请求消耗的总 token 数量（输入 + 输出）。 * **action_usage** `object[]` 本次请求插件用量信息。 * action_usage.**action_name** `string` 插件分类名称，如 `content_plugin`（内容插件）等。 * action_usage.**count** `integer` 本次请求某插件分类插件总调用次数。 * **action_details** `object[]` 本次请求插件调用详情。 * action_details.**name** `string` 插件分类名称，如 `content_plugin`（内容插件）等。 * action_details.**count** `integer` 本次请求某插件分类中插件调用次数。 * action_details.**tool_details** `object[]` 某插件分类中插件调用详细信息。 * action_details.tool_details.**name** `string` 具体调用的工具名称。 * action_details.tool_details.**input** `object` 插件输入参数，调用插件的数据结构。 * action_details.tool_details.**output** `object` 插件输出结果，调用插件返回的数据结构。 * action_details.tool_details.**created_at** `integer` 插件调用开始时间的 Unix 时间戳（秒）。插件调用开始时间。 * action_details.tool_details.**completed_at** `integer` 插件调用结束时间的 Unix 时间戳（秒）。插件调用结束时间。 * **m****etadat****a** `object / null` 额外参数。 * metadata **.****group_chat_config** `object` `必选` 运行时动态传入的角色配置。 * metadata **.** group_chat_config.**characters** `object[]` `必选` 群聊角色列表。 * metadata **.** group_chat_config.characters.**name** `string` `必选` 群聊角色名称。 * metadata **.** group_chat_config.characters.**system_prompt** `string` `必选` 群聊角色设定，用于告知应用需要扮演的角色。 * metadata **.** group_chat_config.characters.**model_desc** `object` `必选` 模型说明。 * metadata **.** group_chat_config.characters.model_desc.**endpoint_id** `string` `必选` 您创建的 [推理接入点](/docs/online-inference-standard#1739de09) ID。 * metadata **.** group_chat_config.**description** `string / null` `默认值 null` 群聊场景描述，可设定群聊的主题、时间地点、事件场景、用户扮演的角色等信息。 * metadata **.** group_chat_config.**user_name** `string / null` `默认值用户` 表示“我”所扮演的角色名称，默认值为“用户”。 * metadata **.****user_info** `string / null` `默认值 null` 可以被反序列化成 json 的字符串，需要包含 `city` 和 `district` 两个字段。 * metadata **.****emit_intention_signal_extra** `string / null` `默认值 "false"` 值是 `"true"` 时，会中途返回intention状态 "正在搜索"。 * metadata **.****target_character_name** `string / null` `默认值 null` 群聊Bot对话时填写，指定本次要发言的角色名，必须是存在于`characters`里的角色。 * **references** `object[]` 本次请求中应用调用了插件的信息。字段参考具体见各个插件的数据结构。 * [联网插件](/docs/online-plugin-data-structures) * [知识库插件](/docs/knowledge-base-plugin-data-structures) **选项：流式调用返回** * **id** `string` 本次请求的唯一标识。 * **model** `string` 本次请求实际使用的模型名称和版本。 doubao 1.5 代模型的模型名称格式为 `doubao-1-5-**`。如调用部署doubao-1.5-pro-32k 250115模型的推理接入点，返回model字段信息doubao-1-5-pro-32k-250115。 * **created** `integer` 本次请求创建时间的 Unix 时间戳（秒）。 * **object** `string` 固定为 `chat.completion.chunk`。 * **choices** `object[]` 本次请求的模型输出内容。 * choices.**index **`integer` 当前元素在 **choices** 列表的索引。 * choices.**finish_reason **`string` 模型停止生成 token 的原因。取值范围： * `stop`：模型输出自然结束，或因命中请求参数 stop 中指定的字段而被截断。 * `length`：模型输出因达到模型输出限制而被截断，有下面几种原因： * 触发`max_token`限制（回答内容的长度限制）。 * 触发`max_completion_tokens`限制（思维链内容+回答内容的长度限制）。 * 触发`context_window` 限制（输入内容+思维链内容+回答内容的长度限制）。 * `content_filter`：模型输出被内容审核拦截。 * `tool_calls`：模型调用了工具。 * choices.**delta **`object` 模型输出的增量内容。 * choices.delta.**role **`string` 内容输出的角色，此处固定为 `assistant`。 * choices.delta.**content **`string` 模型生成的消息内容。 * choices.delta.**reasoning_content **`string / null` 模型处理问题的思维链内容。仅深度推理模型支持返回此字段，深度推理模型请参见 [支持模型](/docs/deep-thinking#fee31637)。 * choices.delta.**tool_calls **`object[] / null` 模型调用的工具列表。 * choices.delta.tool_calls.**i****d **`string` 调用的工具的 ID。 * choices.delta.tool_calls.**index** `integer` 当前元素在模型调用的工具列表的索引。 * choices.delta.tool_calls.**type **`string` 工具类型，当前仅支持`function`。 * choices.delta.tool_calls.**function **`object` 模型调用的函数。 * choices.delta.tool_calls.function.**name **`string` 模型调用的函数的名称。 * choices.delta.tool_calls.function.**arguments **`string` 模型生成的用于调用函数的参数，JSON 格式。模型并不总是生成有效的 JSON，并且可能会虚构出一些您的函数参数规范中未定义的参数。在调用函数之前，请在您的代码中验证这些参数是否有效。 * choices.**logprobs **`object / null` 当前内容的对数概率信息。 * choices.logprobs.**content **`object[] / null` message列表中每个 content 元素中的 token 对数概率信息。 * choices.logprobs.content.**token **`string` 当前 token。 * choices.logprobs.content.**bytes **`integer[] / null` 当前 token 的 UTF-8 值，格式为整数列表。当一个字符由多个 token 组成（表情符号或特殊字符等）时可以用于字符的编码和解码。如果 token 没有 UTF-8 值则为空。 * choices.logprobs.content.**logprob **`float` 当前 token 的对数概率。 * choices.logprobs.content.**top_logprobs **`object[]` 在当前 token 位置最有可能的标记及其对数概率的列表。在一些情况下，返回的数量可能比请求参数 top_logprobs 指定的数量要少。 * choices.logprobs.content.top_logprobs.**token **`string` 当前 token。 * choices.logprobs.content.top_logprobs.**bytes **`integer[] / null` 当前 token 的 UTF-8 值，格式为整数列表。当一个字符由多个 token 组成（表情符号或特殊字符等）时可以用于字符的编码和解码。如果 token 没有 UTF-8 值则为空。 * choices.logprobs.content.top_logprobs.**logprob **`float` 当前 token 的对数概率。 * choices.**moderation_hit_type **`string / null` 模型输出文字含有敏感信息时，会返回模型输出文字命中的风险分类标签。返回值及含义： * `severe_violation`：模型输出文字涉及严重违规。 * `violence`：模型输出文字涉及激进行为。注意：当前只有 [视觉理解模型](/docs/image-understanding#18cf565a) 支持返回该字段，且只有在方舟控制台 [接入点配置页面](https://ark.volcengine.com/region:cn-beijing/endpoint/create?customModelId=) 或者 [CreateEndpoint](/docs/create-endpoint-api) 接口中，将内容护栏方案（ModerationStrategy）设置为基础方案（Basic）时，才会返回风险分类标签。 * **bot_usage** `object` 本次请求的 token 用量。 * bot_usage.**model_usage **`object[]` 本次请求不同 endpoint 的 token 消耗。 * bot_usage.model_usage **.** name `string` 调用的模型 ID。 * bot_usage.model_usage **.** prompt_tokens `integer` 输入的 prompt token 数量，包含用户输入的提示词和调用插件返回的信息。 * bot_usage.model_usage **.** completion_tokens `integer` 模型生成的 token 数量。 * bot_usage.model_usage **.** total_tokens `integer` 本次请求消耗的总 token 数量（输入 + 输出）。 * action_usage `object[]` 本次请求插件用量信息。 * action_usage.action_name `string` 插件分类名称，如 `content_plugin`（内容插件）等。 * action_usage.count `integer` 本次请求某插件分类插件总调用次数。 * action_details `object[]` 本次请求插件调用详情。 * action_details.name `string` 插件分类名称，如 `content_plugin`（内容插件）等。 * action_details.count `integer` 本次请求某插件分类中插件调用次数。 * action_details.tool_details `object[]` 某插件分类中插件调用详细信息。 * action_details.tool_details.name `string` 具体调用的工具名称。 * action_details.tool_details.input `object` 插件输入参数，调用插件的数据结构。 * action_details.tool_details.output `object` 插件输出结果，调用插件返回的数据结构。 * action_details.tool_details.created_at `integer` 插件调用开始时间。 * action_details.tool_details.completed_at `integer` 插件调用结束时间。 * **m****etadat****a** `object / null` 额外参数。 * metadata **.****group_chat_config** `object` `必选` 运行时动态传入的角色配置。 * metadata **.** group_chat_config.**characters** `object[]` `必选` 群聊角色列表。 * metadata **.** group_chat_config.characters.**name** `string` `必选` 群聊角色名称。 * metadata **.** group_chat_config.characters.**system_prompt** `string` `必选` 群聊角色设定，用于告知应用需要扮演的角色。 * metadata **.** group_chat_config.characters.**model_desc** `object` `必选` 模型说明。 * metadata **.** group_chat_config.characters.model_desc.**endpoint_id** `string` `必选` 您创建的 [推理接入点](/docs/online-inference-standard#1739de09) ID。 * metadata **.** group_chat_config.**description** `string / null` `默认值 null` 群聊场景描述，可设定群聊的主题、时间地点、事件场景、用户扮演的角色等信息。 * metadata **.** group_chat_config.**user_name** `string / null` `默认值用户` 表示“我”所扮演的角色名称，默认值为“用户”。 * metadata **.****user_info** `string / null` `默认值 null` 可以被反序列化成 json 的字符串，需要包含 `city` 和 `district` 两个字段。 * metadata **.****emit_intention_signal_extra** `string / null` `默认值 "false"` 值是 `"true"` 时，会中途返回intention状态 "正在搜索"。 * metadata **.****target_character_name** `string / null` `默认值 null` 群聊Bot对话时填写，指定本次要发言的角色名，必须是存在于`characters`里的角色。 * **references** `object[]` 本次请求中应用调用了插件的信息。字段参考具体见各个插件的数据结构。 * [联网插件](/docs/online-plugin-data-structures) * [知识库插件](/docs/knowledge-base-plugin-data-structures) ## 代码示例 ## 联网应用 ### cURL #### Request ```bash curl https://ark.cn-beijing.volces.com/api/v3/bots/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $ARK_API_KEY" \ -d $'{ "messages": [ { "content": "杭州今天天气", "role": "user" } ], "model": "bot-20250407225237-6xv7r" }' ``` #### Response ``` { "id": "02174410213780126458cc51545801ff6ffb59c94a55dae86c43b", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "以下是杭州今天（2025年4月8日）的天气情况：\n- **整体天气状况**：今天多云，偏南风3级。\n- **气温情况**：白天最高气温31度，平均相对湿度55%。\n- **空气质量**：杭州市区AQI为90 - 110，首要污染物是O₃，空气质量等级为良 - 轻度污染，易感人群需减少户外锻炼。 ", "role": "assistant" } } ], "created": 1744102143, "model": "doubao-1-5-pro-32k-250115", "object": "chat.completion", "bot_usage": { "model_usage": [ { "name": "doubao-1-5-pro-32k-250115", "prompt_tokens": 1960, "completion_tokens": 99, "total_tokens": 2059 } ], "action_usage": [ { "action_name": "content_plugin", "count": 1 } ], "action_details": [ { "name": "content_plugin", "count": 1, "tool_details": [ { "name": "search", "input": { "queries": [ "杭州今天天气" ], "source_type": [ "search_engine" ], "offset": 0, "count": 5, "account_id": 2100000825, "video_token": "" }, "output": { "type": "tool", "data": { "status_code": 200, "data": { "has_more": true, "results": [ { "id": "3ba8807eabf8793d-65e575422f8ebfb2", "source_type": "search_engine", "site_name": "搜索引擎-中国天气", "title": "【杭州天气预报15天_杭州天气预报15天查询】-中国天气网", "summary": "20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级 20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级", "publish_time": 0, "url": "https://m.weather.com.cn/mweather15d/101210101.shtml#!/abighis", "cover_image": null, "search_plugin_data": { "doc_id": "3ba8807eabf8793d-65e575422f8ebfb2", "logo_url": "https://p3-search.byteimg.com/img/labis/dafd663cfa3b7fce9addcca7916010cb~noop.jpeg", "auth_info": "非常权威", "auth_score": 1, "rel_info": "强相关", "rel_score": 0.994623, "freshness_info": "非常满足时效需求", "freshness_score": 0.994948, "final_ref": "", "final_ref_len": 0, "author_name": "", "duration": 0 }, "smart_content_data": null, "ruyi_data": null }, ... { "id": "d4d70607487f997f-b2adf6b6d837f126", "source_type": "search_engine", "site_name": "搜索引擎-www.hzqx.com", "title": "杭州天气", "summary": "杭州国家基准气候站发布 25.7℃ 较昨日同期高2.8℃ 相对湿度：40 % 风力：东南2级气压： 1012.1 hPa 近1小时降水:0 mm 能见度：22652 m *实时数据、未经人工复核杭州市气象台4月8日8时50分发布的市区天气预报：\n今天多云；明天多云转阴，傍晚到夜里局部有阵雨；后天阴转多云。今天偏南风3级。今天白天最高气温31度，明天白天最高气温32度，明天早晨最低气温21度，今天平均相对湿度55%。2025年04月08日 08时50分过去24小时主城区未来7天天气客观预报（该预报数据为数值预报自动输出，未经人工订正，仅供参考）白天 08时 | 20时夜晚昨天20时 | 当天08时今天明天周四周五周六周日周一 04/08 04/09 04/10 04/11 04/12 04/13 04/14 多云多云晴多云多云晴晴晴多云晴大雨晴晴东风东南风东北风东风北风西南风东北风不超过3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 31.7℃ 32.2℃ 32.2℃ 31.2℃ 19.4℃ 25.2℃ 27.3℃ 17.6℃ 19.4℃ 21.5℃ 19.3℃ 17.7℃ 11.2℃ 14.7℃ * 白天为当天08时到20时，通常气温为最高温度；夜晚为昨天20时到当天08时，通常气温为最低温度。杭州市当前气温2025-04-08 10:30 分布图今日生活气象空气质量预报杭州市环境监测站和杭州市气象台联合发布 4月8日杭州市区AQI:90-110，首要污染物:O3，空气质量等级:良-轻度污染。易感人群减少户外锻炼。4月9日杭州市区AQI:80-100，首要污染物:O3，空气质量等级:良。", "publish_time": 1744079400, "url": "https://www.hzqx.com/hztq/index.html", "cover_image": null, "search_plugin_data": { "doc_id": "d4d70607487f997f-b2adf6b6d837f126", "logo_url": "", "auth_info": "非常权威", "auth_score": 0.875495, "rel_info": "强相关", "rel_score": 0.834824, "freshness_info": "非常满足时效需求", "freshness_score": 0.994948, "final_ref": "", "final_ref_len": 0, "author_name": "", "duration": 0 }, "smart_content_data": null, "ruyi_data": null } ], "debug_data": { "extra": null }, "log_id": "02174410213780126458cc51545801ff6ffb59c94a55dae86c43b", "timing": { "total": 1431 } } } }, "created_at": 1744102137810, "completed_at": 1744102139264 } ] } ] }, "metadata": {}, "references": [ { "url": "https://m.weather.com.cn/mweather15d/101210101.shtml#!/abighis", "logo_url": "https://p3-search.byteimg.com/img/labis/dafd663cfa3b7fce9addcca7916010cb~noop.jpeg", "site_name": "搜索引擎-中国天气", "title": "【杭州天气预报15天_杭州天气预报15天查询】-中国天气网", "summary": "20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级 20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级", "publish_time": "1970年01月01日 08:00:00(CST) 星期四", "extra": { "rel_info": "强相关", "freshness_info": "非常满足时效需求", "auth_info": "非常权威", "final_ref": "" } }, ... { "url": "https://www.hzqx.com/hztq/index.html", "logo_url": "", "site_name": "搜索引擎-www.hzqx.com", "title": "杭州天气", "summary": "杭州国家基准气候站发布 25.7℃ 较昨日同期高2.8℃ 相对湿度：40 % 风力：东南2级气压： 1012.1 hPa 近1小时降水:0 mm 能见度：22652 m *实时数据、未经人工复核杭州市气象台4月8日8时50分发布的市区天气预报：\n今天多云；明天多云转阴，傍晚到夜里局部有阵雨；后天阴转多云。今天偏南风3级。今天白天最高气温31度，明天白天最高气温32度，明天早晨最低气温21度，今天平均相对湿度55%。2025年04月08日 08时50分过去24小时主城区未来7天天气客观预报（该预报数据为数值预报自动输出，未经人工订正，仅供参考）白天 08时 | 20时夜晚昨天20时 | 当天08时今天明天周四周五周六周日周一 04/08 04/09 04/10 04/11 04/12 04/13 04/14 多云多云晴多云多云晴晴晴多云晴大雨晴晴东风东南风东北风东风北风西南风东北风不超过3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 31.7℃ 32.2℃ 32.2℃ 31.2℃ 19.4℃ 25.2℃ 27.3℃ 17.6℃ 19.4℃ 21.5℃ 19.3℃ 17.7℃ 11.2℃ 14.7℃ * 白天为当天08时到20时，通常气温为最高温度；夜晚为昨天20时到当天08时，通常气温为最低温度。杭州市当前气温2025-04-08 10:30 分布图今日生活气象空气质量预报杭州市环境监测站和杭州市气象台联合发布 4月8日杭州市区AQI:90-110，首要污染物:O3，空气质量等级:良-轻度污染。易感人群减少户外锻炼。4月9日杭州市区AQI:80-100，首要污染物:O3，空气质量等级:良。", "publish_time": "2025年04月08日 10:30:00(CST) 星期二", "extra": { "rel_info": "强相关", "freshness_info": "非常满足时效需求", "auth_info": "非常权威", "final_ref": "" } } ] } ``` ### Python #### Request ```python import os from volcenginesdkarkruntime import Ark client = Ark(api_key=os.environ.get("ARK_API_KEY")) if __name__ == "__main__": resp = client.bot_chat.completions.create( model="bot-20250407******-*****", messages=[{"content":"杭州今天天气","role":"user"}], ) print(resp.choices[0].message.content) ``` #### Response ``` { "id": "02174410213780126458cc51545801ff6ffb59c94a55dae86c43b", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "以下是杭州今天（2025年4月8日）的天气情况：\n- **整体天气状况**：今天多云，偏南风3级。\n- **气温情况**：白天最高气温31度，平均相对湿度55%。\n- **空气质量**：杭州市区AQI为90 - 110，首要污染物是O₃，空气质量等级为良 - 轻度污染，易感人群需减少户外锻炼。 ", "role": "assistant" } } ], "created": 1744102143, "model": "doubao-1-5-pro-32k-250115", "object": "chat.completion", "bot_usage": { "model_usage": [ { "name": "doubao-1-5-pro-32k-250115", "prompt_tokens": 1960, "completion_tokens": 99, "total_tokens": 2059 } ], "action_usage": [ { "action_name": "content_plugin", "count": 1 } ], "action_details": [ { "name": "content_plugin", "count": 1, "tool_details": [ { "name": "search", "input": { "queries": [ "杭州今天天气" ], "source_type": [ "search_engine" ], "offset": 0, "count": 5, "account_id": 2100000825, "video_token": "" }, "output": { "type": "tool", "data": { "status_code": 200, "data": { "has_more": true, "results": [ { "id": "3ba8807eabf8793d-65e575422f8ebfb2", "source_type": "search_engine", "site_name": "搜索引擎-中国天气", "title": "【杭州天气预报15天_杭州天气预报15天查询】-中国天气网", "summary": "20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级 20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级", "publish_time": 0, "url": "https://m.weather.com.cn/mweather15d/101210101.shtml#!/abighis", "cover_image": null, "search_plugin_data": { "doc_id": "3ba8807eabf8793d-65e575422f8ebfb2", "logo_url": "https://p3-search.byteimg.com/img/labis/dafd663cfa3b7fce9addcca7916010cb~noop.jpeg", "auth_info": "非常权威", "auth_score": 1, "rel_info": "强相关", "rel_score": 0.994623, "freshness_info": "非常满足时效需求", "freshness_score": 0.994948, "final_ref": "", "final_ref_len": 0, "author_name": "", "duration": 0 }, "smart_content_data": null, "ruyi_data": null }, ... { "id": "d4d70607487f997f-b2adf6b6d837f126", "source_type": "search_engine", "site_name": "搜索引擎-www.hzqx.com", "title": "杭州天气", "summary": "杭州国家基准气候站发布 25.7℃ 较昨日同期高2.8℃ 相对湿度：40 % 风力：东南2级气压： 1012.1 hPa 近1小时降水:0 mm 能见度：22652 m *实时数据、未经人工复核杭州市气象台4月8日8时50分发布的市区天气预报：\n今天多云；明天多云转阴，傍晚到夜里局部有阵雨；后天阴转多云。今天偏南风3级。今天白天最高气温31度，明天白天最高气温32度，明天早晨最低气温21度，今天平均相对湿度55%。2025年04月08日 08时50分过去24小时主城区未来7天天气客观预报（该预报数据为数值预报自动输出，未经人工订正，仅供参考）白天 08时 | 20时夜晚昨天20时 | 当天08时今天明天周四周五周六周日周一 04/08 04/09 04/10 04/11 04/12 04/13 04/14 多云多云晴多云多云晴晴晴多云晴大雨晴晴东风东南风东北风东风北风西南风东北风不超过3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 31.7℃ 32.2℃ 32.2℃ 31.2℃ 19.4℃ 25.2℃ 27.3℃ 17.6℃ 19.4℃ 21.5℃ 19.3℃ 17.7℃ 11.2℃ 14.7℃ * 白天为当天08时到20时，通常气温为最高温度；夜晚为昨天20时到当天08时，通常气温为最低温度。杭州市当前气温2025-04-08 10:30 分布图今日生活气象空气质量预报杭州市环境监测站和杭州市气象台联合发布 4月8日杭州市区AQI:90-110，首要污染物:O3，空气质量等级:良-轻度污染。易感人群减少户外锻炼。4月9日杭州市区AQI:80-100，首要污染物:O3，空气质量等级:良。", "publish_time": 1744079400, "url": "https://www.hzqx.com/hztq/index.html", "cover_image": null, "search_plugin_data": { "doc_id": "d4d70607487f997f-b2adf6b6d837f126", "logo_url": "", "auth_info": "非常权威", "auth_score": 0.875495, "rel_info": "强相关", "rel_score": 0.834824, "freshness_info": "非常满足时效需求", "freshness_score": 0.994948, "final_ref": "", "final_ref_len": 0, "author_name": "", "duration": 0 }, "smart_content_data": null, "ruyi_data": null } ], "debug_data": { "extra": null }, "log_id": "02174410213780126458cc51545801ff6ffb59c94a55dae86c43b", "timing": { "total": 1431 } } } }, "created_at": 1744102137810, "completed_at": 1744102139264 } ] } ] }, "metadata": {}, "references": [ { "url": "https://m.weather.com.cn/mweather15d/101210101.shtml#!/abighis", "logo_url": "https://p3-search.byteimg.com/img/labis/dafd663cfa3b7fce9addcca7916010cb~noop.jpeg", "site_name": "搜索引擎-中国天气", "title": "【杭州天气预报15天_杭州天气预报15天查询】-中国天气网", "summary": "20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级 20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级", "publish_time": "1970年01月01日 08:00:00(CST) 星期四", "extra": { "rel_info": "强相关", "freshness_info": "非常满足时效需求", "auth_info": "非常权威", "final_ref": "" } }, ... { "url": "https://www.hzqx.com/hztq/index.html", "logo_url": "", "site_name": "搜索引擎-www.hzqx.com", "title": "杭州天气", "summary": "杭州国家基准气候站发布 25.7℃ 较昨日同期高2.8℃ 相对湿度：40 % 风力：东南2级气压： 1012.1 hPa 近1小时降水:0 mm 能见度：22652 m *实时数据、未经人工复核杭州市气象台4月8日8时50分发布的市区天气预报：\n今天多云；明天多云转阴，傍晚到夜里局部有阵雨；后天阴转多云。今天偏南风3级。今天白天最高气温31度，明天白天最高气温32度，明天早晨最低气温21度，今天平均相对湿度55%。2025年04月08日 08时50分过去24小时主城区未来7天天气客观预报（该预报数据为数值预报自动输出，未经人工订正，仅供参考）白天 08时 | 20时夜晚昨天20时 | 当天08时今天明天周四周五周六周日周一 04/08 04/09 04/10 04/11 04/12 04/13 04/14 多云多云晴多云多云晴晴晴多云晴大雨晴晴东风东南风东北风东风北风西南风东北风不超过3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 31.7℃ 32.2℃ 32.2℃ 31.2℃ 19.4℃ 25.2℃ 27.3℃ 17.6℃ 19.4℃ 21.5℃ 19.3℃ 17.7℃ 11.2℃ 14.7℃ * 白天为当天08时到20时，通常气温为最高温度；夜晚为昨天20时到当天08时，通常气温为最低温度。杭州市当前气温2025-04-08 10:30 分布图今日生活气象空气质量预报杭州市环境监测站和杭州市气象台联合发布 4月8日杭州市区AQI:90-110，首要污染物:O3，空气质量等级:良-轻度污染。易感人群减少户外锻炼。4月9日杭州市区AQI:80-100，首要污染物:O3，空气质量等级:良。", "publish_time": "2025年04月08日 10:30:00(CST) 星期二", "extra": { "rel_info": "强相关", "freshness_info": "非常满足时效需求", "auth_info": "非常权威", "final_ref": "" } } ] } ``` ### Go #### Request ```go package main import ( "context" "fmt" "os" "github.com/volcengine/volcengine-go-sdk/service/arkruntime" "github.com/volcengine/volcengine-go-sdk/service/arkruntime/model" "github.com/volcengine/volcengine-go-sdk/volcengine" ) func main() { client := arkruntime.NewClientWithApiKey(os.Getenv("ARK_API_KEY")) ctx := context.Background() req := model.BotChatCompletionRequest{ Model: "bot-20250407******-*****", Messages: []*model.ChatCompletionMessage{ &model.ChatCompletionMessage{ Role: "user", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("杭州今天天气"), }, }, }, } fmt.Println("----- standard request -----") resp, err := client.CreateBotChatCompletion(ctx, req) if err != nil { fmt.Printf("standard chat error: %v\n", err) return } if resp.References != nil { for _, ref := range resp.References { fmt.Printf("reference url: %s\n", ref.Url) } } } ``` #### Response ``` { "id": "02174410213780126458cc51545801ff6ffb59c94a55dae86c43b", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "以下是杭州今天（2025年4月8日）的天气情况：\n- **整体天气状况**：今天多云，偏南风3级。\n- **气温情况**：白天最高气温31度，平均相对湿度55%。\n- **空气质量**：杭州市区AQI为90 - 110，首要污染物是O₃，空气质量等级为良 - 轻度污染，易感人群需减少户外锻炼。 ", "role": "assistant" } } ], "created": 1744102143, "model": "doubao-1-5-pro-32k-250115", "object": "chat.completion", "bot_usage": { "model_usage": [ { "name": "doubao-1-5-pro-32k-250115", "prompt_tokens": 1960, "completion_tokens": 99, "total_tokens": 2059 } ], "action_usage": [ { "action_name": "content_plugin", "count": 1 } ], "action_details": [ { "name": "content_plugin", "count": 1, "tool_details": [ { "name": "search", "input": { "queries": [ "杭州今天天气" ], "source_type": [ "search_engine" ], "offset": 0, "count": 5, "account_id": 2100000825, "video_token": "" }, "output": { "type": "tool", "data": { "status_code": 200, "data": { "has_more": true, "results": [ { "id": "3ba8807eabf8793d-65e575422f8ebfb2", "source_type": "search_engine", "site_name": "搜索引擎-中国天气", "title": "【杭州天气预报15天_杭州天气预报15天查询】-中国天气网", "summary": "20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级 20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级", "publish_time": 0, "url": "https://m.weather.com.cn/mweather15d/101210101.shtml#!/abighis", "cover_image": null, "search_plugin_data": { "doc_id": "3ba8807eabf8793d-65e575422f8ebfb2", "logo_url": "https://p3-search.byteimg.com/img/labis/dafd663cfa3b7fce9addcca7916010cb~noop.jpeg", "auth_info": "非常权威", "auth_score": 1, "rel_info": "强相关", "rel_score": 0.994623, "freshness_info": "非常满足时效需求", "freshness_score": 0.994948, "final_ref": "", "final_ref_len": 0, "author_name": "", "duration": 0 }, "smart_content_data": null, "ruyi_data": null }, ... { "id": "d4d70607487f997f-b2adf6b6d837f126", "source_type": "search_engine", "site_name": "搜索引擎-www.hzqx.com", "title": "杭州天气", "summary": "杭州国家基准气候站发布 25.7℃ 较昨日同期高2.8℃ 相对湿度：40 % 风力：东南2级气压： 1012.1 hPa 近1小时降水:0 mm 能见度：22652 m *实时数据、未经人工复核杭州市气象台4月8日8时50分发布的市区天气预报：\n今天多云；明天多云转阴，傍晚到夜里局部有阵雨；后天阴转多云。今天偏南风3级。今天白天最高气温31度，明天白天最高气温32度，明天早晨最低气温21度，今天平均相对湿度55%。2025年04月08日 08时50分过去24小时主城区未来7天天气客观预报（该预报数据为数值预报自动输出，未经人工订正，仅供参考）白天 08时 | 20时夜晚昨天20时 | 当天08时今天明天周四周五周六周日周一 04/08 04/09 04/10 04/11 04/12 04/13 04/14 多云多云晴多云多云晴晴晴多云晴大雨晴晴东风东南风东北风东风北风西南风东北风不超过3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 31.7℃ 32.2℃ 32.2℃ 31.2℃ 19.4℃ 25.2℃ 27.3℃ 17.6℃ 19.4℃ 21.5℃ 19.3℃ 17.7℃ 11.2℃ 14.7℃ * 白天为当天08时到20时，通常气温为最高温度；夜晚为昨天20时到当天08时，通常气温为最低温度。杭州市当前气温2025-04-08 10:30 分布图今日生活气象空气质量预报杭州市环境监测站和杭州市气象台联合发布 4月8日杭州市区AQI:90-110，首要污染物:O3，空气质量等级:良-轻度污染。易感人群减少户外锻炼。4月9日杭州市区AQI:80-100，首要污染物:O3，空气质量等级:良。", "publish_time": 1744079400, "url": "https://www.hzqx.com/hztq/index.html", "cover_image": null, "search_plugin_data": { "doc_id": "d4d70607487f997f-b2adf6b6d837f126", "logo_url": "", "auth_info": "非常权威", "auth_score": 0.875495, "rel_info": "强相关", "rel_score": 0.834824, "freshness_info": "非常满足时效需求", "freshness_score": 0.994948, "final_ref": "", "final_ref_len": 0, "author_name": "", "duration": 0 }, "smart_content_data": null, "ruyi_data": null } ], "debug_data": { "extra": null }, "log_id": "02174410213780126458cc51545801ff6ffb59c94a55dae86c43b", "timing": { "total": 1431 } } } }, "created_at": 1744102137810, "completed_at": 1744102139264 } ] } ] }, "metadata": {}, "references": [ { "url": "https://m.weather.com.cn/mweather15d/101210101.shtml#!/abighis", "logo_url": "https://p3-search.byteimg.com/img/labis/dafd663cfa3b7fce9addcca7916010cb~noop.jpeg", "site_name": "搜索引擎-中国天气", "title": "【杭州天气预报15天_杭州天气预报15天查询】-中国天气网", "summary": "20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级 20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级", "publish_time": "1970年01月01日 08:00:00(CST) 星期四", "extra": { "rel_info": "强相关", "freshness_info": "非常满足时效需求", "auth_info": "非常权威", "final_ref": "" } }, ... { "url": "https://www.hzqx.com/hztq/index.html", "logo_url": "", "site_name": "搜索引擎-www.hzqx.com", "title": "杭州天气", "summary": "杭州国家基准气候站发布 25.7℃ 较昨日同期高2.8℃ 相对湿度：40 % 风力：东南2级气压： 1012.1 hPa 近1小时降水:0 mm 能见度：22652 m *实时数据、未经人工复核杭州市气象台4月8日8时50分发布的市区天气预报：\n今天多云；明天多云转阴，傍晚到夜里局部有阵雨；后天阴转多云。今天偏南风3级。今天白天最高气温31度，明天白天最高气温32度，明天早晨最低气温21度，今天平均相对湿度55%。2025年04月08日 08时50分过去24小时主城区未来7天天气客观预报（该预报数据为数值预报自动输出，未经人工订正，仅供参考）白天 08时 | 20时夜晚昨天20时 | 当天08时今天明天周四周五周六周日周一 04/08 04/09 04/10 04/11 04/12 04/13 04/14 多云多云晴多云多云晴晴晴多云晴大雨晴晴东风东南风东北风东风北风西南风东北风不超过3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 31.7℃ 32.2℃ 32.2℃ 31.2℃ 19.4℃ 25.2℃ 27.3℃ 17.6℃ 19.4℃ 21.5℃ 19.3℃ 17.7℃ 11.2℃ 14.7℃ * 白天为当天08时到20时，通常气温为最高温度；夜晚为昨天20时到当天08时，通常气温为最低温度。杭州市当前气温2025-04-08 10:30 分布图今日生活气象空气质量预报杭州市环境监测站和杭州市气象台联合发布 4月8日杭州市区AQI:90-110，首要污染物:O3，空气质量等级:良-轻度污染。易感人群减少户外锻炼。4月9日杭州市区AQI:80-100，首要污染物:O3，空气质量等级:良。", "publish_time": "2025年04月08日 10:30:00(CST) 星期二", "extra": { "rel_info": "强相关", "freshness_info": "非常满足时效需求", "auth_info": "非常权威", "final_ref": "" } } ] } ``` ### Java #### Request ```java package com.volcengine.sample; import com.fasterxml.jackson.core.JsonProcessingException; import com.volcengine.ark.runtime.model.bot.completion.chat.BotChatCompletionRequest; import com.volcengine.ark.runtime.model.completion.chat.*; import com.volcengine.ark.runtime.model.completion.chat.ChatCompletionRequest.*; import com.volcengine.ark.runtime.service.ArkService; import java.util.ArrayList; import java.util.List; import java.util.concurrent.TimeUnit; import okhttp3.ConnectionPool; import okhttp3.Dispatcher; /* # pom.xml com.volcengine volcengine-java-sdk-ark-runtime // 替换正式版本号 LATEST */ public class Sample { static String apiKey = System.getenv("ARK_API_KEY"); static ConnectionPool connectionPool = new ConnectionPool(5, 1, TimeUnit.SECONDS); static Dispatcher dispatcher = new Dispatcher(); static ArkService service = ArkService.builder() .dispatcher(dispatcher) .connectionPool(connectionPool) .apiKey(apiKey) .baseUrl("https://ark.cn-beijing.volces.com/api/v3") .build(); public static void main(String[] args) throws JsonProcessingException { List messagesForReqList = new ArrayList<>(); ChatMessage elementForMessagesForReqList0 = ChatMessage.builder().role(ChatMessageRole.USER).content("杭州今天天气").build(); messagesForReqList.add(elementForMessagesForReqList0); BotChatCompletionRequest req = BotChatCompletionRequest.builder() .model("bot-20250407******-*****") .messages(messagesForReqList) .build(); service.createBotChatCompletion(req) .getChoices() .forEach(choice -> System.out.println(choice.getMessage().getContent())); // shutdown service after all requests is finished service.shutdownExecutor(); } } ``` #### Response ``` { "id": "02174410213780126458cc51545801ff6ffb59c94a55dae86c43b", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "以下是杭州今天（2025年4月8日）的天气情况：\n- **整体天气状况**：今天多云，偏南风3级。\n- **气温情况**：白天最高气温31度，平均相对湿度55%。\n- **空气质量**：杭州市区AQI为90 - 110，首要污染物是O₃，空气质量等级为良 - 轻度污染，易感人群需减少户外锻炼。 ", "role": "assistant" } } ], "created": 1744102143, "model": "doubao-1-5-pro-32k-250115", "object": "chat.completion", "bot_usage": { "model_usage": [ { "name": "doubao-1-5-pro-32k-250115", "prompt_tokens": 1960, "completion_tokens": 99, "total_tokens": 2059 } ], "action_usage": [ { "action_name": "content_plugin", "count": 1 } ], "action_details": [ { "name": "content_plugin", "count": 1, "tool_details": [ { "name": "search", "input": { "queries": [ "杭州今天天气" ], "source_type": [ "search_engine" ], "offset": 0, "count": 5, "account_id": 2100000825, "video_token": "" }, "output": { "type": "tool", "data": { "status_code": 200, "data": { "has_more": true, "results": [ { "id": "3ba8807eabf8793d-65e575422f8ebfb2", "source_type": "search_engine", "site_name": "搜索引擎-中国天气", "title": "【杭州天气预报15天_杭州天气预报15天查询】-中国天气网", "summary": "20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级 20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级", "publish_time": 0, "url": "https://m.weather.com.cn/mweather15d/101210101.shtml#!/abighis", "cover_image": null, "search_plugin_data": { "doc_id": "3ba8807eabf8793d-65e575422f8ebfb2", "logo_url": "https://p3-search.byteimg.com/img/labis/dafd663cfa3b7fce9addcca7916010cb~noop.jpeg", "auth_info": "非常权威", "auth_score": 1, "rel_info": "强相关", "rel_score": 0.994623, "freshness_info": "非常满足时效需求", "freshness_score": 0.994948, "final_ref": "", "final_ref_len": 0, "author_name": "", "duration": 0 }, "smart_content_data": null, "ruyi_data": null }, ... { "id": "d4d70607487f997f-b2adf6b6d837f126", "source_type": "search_engine", "site_name": "搜索引擎-www.hzqx.com", "title": "杭州天气", "summary": "杭州国家基准气候站发布 25.7℃ 较昨日同期高2.8℃ 相对湿度：40 % 风力：东南2级气压： 1012.1 hPa 近1小时降水:0 mm 能见度：22652 m *实时数据、未经人工复核杭州市气象台4月8日8时50分发布的市区天气预报：\n今天多云；明天多云转阴，傍晚到夜里局部有阵雨；后天阴转多云。今天偏南风3级。今天白天最高气温31度，明天白天最高气温32度，明天早晨最低气温21度，今天平均相对湿度55%。2025年04月08日 08时50分过去24小时主城区未来7天天气客观预报（该预报数据为数值预报自动输出，未经人工订正，仅供参考）白天 08时 | 20时夜晚昨天20时 | 当天08时今天明天周四周五周六周日周一 04/08 04/09 04/10 04/11 04/12 04/13 04/14 多云多云晴多云多云晴晴晴多云晴大雨晴晴东风东南风东北风东风北风西南风东北风不超过3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 31.7℃ 32.2℃ 32.2℃ 31.2℃ 19.4℃ 25.2℃ 27.3℃ 17.6℃ 19.4℃ 21.5℃ 19.3℃ 17.7℃ 11.2℃ 14.7℃ * 白天为当天08时到20时，通常气温为最高温度；夜晚为昨天20时到当天08时，通常气温为最低温度。杭州市当前气温2025-04-08 10:30 分布图今日生活气象空气质量预报杭州市环境监测站和杭州市气象台联合发布 4月8日杭州市区AQI:90-110，首要污染物:O3，空气质量等级:良-轻度污染。易感人群减少户外锻炼。4月9日杭州市区AQI:80-100，首要污染物:O3，空气质量等级:良。", "publish_time": 1744079400, "url": "https://www.hzqx.com/hztq/index.html", "cover_image": null, "search_plugin_data": { "doc_id": "d4d70607487f997f-b2adf6b6d837f126", "logo_url": "", "auth_info": "非常权威", "auth_score": 0.875495, "rel_info": "强相关", "rel_score": 0.834824, "freshness_info": "非常满足时效需求", "freshness_score": 0.994948, "final_ref": "", "final_ref_len": 0, "author_name": "", "duration": 0 }, "smart_content_data": null, "ruyi_data": null } ], "debug_data": { "extra": null }, "log_id": "02174410213780126458cc51545801ff6ffb59c94a55dae86c43b", "timing": { "total": 1431 } } } }, "created_at": 1744102137810, "completed_at": 1744102139264 } ] } ] }, "metadata": {}, "references": [ { "url": "https://m.weather.com.cn/mweather15d/101210101.shtml#!/abighis", "logo_url": "https://p3-search.byteimg.com/img/labis/dafd663cfa3b7fce9addcca7916010cb~noop.jpeg", "site_name": "搜索引擎-中国天气", "title": "【杭州天气预报15天_杭州天气预报15天查询】-中国天气网", "summary": "20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级 20/8℃ 东北风<3级北风<3级 19/7℃ 东风<3级南风<3级 26/12℃ 东风<3级西风<3级", "publish_time": "1970年01月01日 08:00:00(CST) 星期四", "extra": { "rel_info": "强相关", "freshness_info": "非常满足时效需求", "auth_info": "非常权威", "final_ref": "" } }, ... { "url": "https://www.hzqx.com/hztq/index.html", "logo_url": "", "site_name": "搜索引擎-www.hzqx.com", "title": "杭州天气", "summary": "杭州国家基准气候站发布 25.7℃ 较昨日同期高2.8℃ 相对湿度：40 % 风力：东南2级气压： 1012.1 hPa 近1小时降水:0 mm 能见度：22652 m *实时数据、未经人工复核杭州市气象台4月8日8时50分发布的市区天气预报：\n今天多云；明天多云转阴，傍晚到夜里局部有阵雨；后天阴转多云。今天偏南风3级。今天白天最高气温31度，明天白天最高气温32度，明天早晨最低气温21度，今天平均相对湿度55%。2025年04月08日 08时50分过去24小时主城区未来7天天气客观预报（该预报数据为数值预报自动输出，未经人工订正，仅供参考）白天 08时 | 20时夜晚昨天20时 | 当天08时今天明天周四周五周六周日周一 04/08 04/09 04/10 04/11 04/12 04/13 04/14 多云多云晴多云多云晴晴晴多云晴大雨晴晴东风东南风东北风东风北风西南风东北风不超过3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 ≤3级 31.7℃ 32.2℃ 32.2℃ 31.2℃ 19.4℃ 25.2℃ 27.3℃ 17.6℃ 19.4℃ 21.5℃ 19.3℃ 17.7℃ 11.2℃ 14.7℃ * 白天为当天08时到20时，通常气温为最高温度；夜晚为昨天20时到当天08时，通常气温为最低温度。杭州市当前气温2025-04-08 10:30 分布图今日生活气象空气质量预报杭州市环境监测站和杭州市气象台联合发布 4月8日杭州市区AQI:90-110，首要污染物:O3，空气质量等级:良-轻度污染。易感人群减少户外锻炼。4月9日杭州市区AQI:80-100，首要污染物:O3，空气质量等级:良。", "publish_time": "2025年04月08日 10:30:00(CST) 星期二", "extra": { "rel_info": "强相关", "freshness_info": "非常满足时效需求", "auth_info": "非常权威", "final_ref": "" } } ] } ``` ## 单轮对话 ### cURL #### Request ```bash curl 'https://ark.cn-beijing.volces.com/api/v3/bots/chat/completions' \ -H "Authorization: Bearer $ARK_API_KEY" \ -H 'Content-Type: application/json' \ -d '{ "model": "bot-20250707202501-k6czp", "messages": [ { "role": "system", "content": "You are a helpful assistant." }, { "role": "user", "content": "Hello!" } ] }' ``` #### Response ``` { "id": "021751891462122d1a8c9bcb9c54eb50913f5e75c28a3ac724183", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "\nHello there! How can I assist you today? 😊", "role": "assistant", "reasoning_content": "Hmm, the user just said \"Hello!\" - a simple and friendly greeting. Since this is likely the start of a conversation, they're probably testing the waters or initiating casual interaction. No complex needs here, but I should match their upbeat tone. \n\nAh, the response was straightforward - \"Hello there! How can I assist you today?\" Perfectly neutral yet warm opening. Kept it open-ended to invite follow-up. The exclamation point adds energy without being overbearing. \n\nWonder if they're new or returning? Either way, no need to overthink - just acknowledge and pivot to assistance. Short responses like this usually mean they'll either ask something specific next or vanish. Better keep the reply crisp.\n" } } ], "created": 1751891468, "model": "deepseek-r1-250528", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 35, "completion_tokens": 159, "total_tokens": 194 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ### Python #### Request ```python import os from volcenginesdkarkruntime import Ark # For more information，please check this document（https://ark.volcengine.com/region:cn-beijing/docs/api-key） client = Ark(api_key=os.environ.get("ARK_API_KEY")) if __name__ == "__main__": resp = client.bot_chat.completions.create( model="bot-20250707202501-k6czp", messages=[{"content":"You are a helpful assistant.","role":"system"},{"content":"Hello!","role":"user"}], ) print(resp.choices[0].message.content) ``` #### Response ``` { "id": "021751891462122d1a8c9bcb9c54eb50913f5e75c28a3ac724183", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "\nHello there! How can I assist you today? 😊", "role": "assistant", "reasoning_content": "Hmm, the user just said \"Hello!\" - a simple and friendly greeting. Since this is likely the start of a conversation, they're probably testing the waters or initiating casual interaction. No complex needs here, but I should match their upbeat tone. \n\nAh, the response was straightforward - \"Hello there! How can I assist you today?\" Perfectly neutral yet warm opening. Kept it open-ended to invite follow-up. The exclamation point adds energy without being overbearing. \n\nWonder if they're new or returning? Either way, no need to overthink - just acknowledge and pivot to assistance. Short responses like this usually mean they'll either ask something specific next or vanish. Better keep the reply crisp.\n" } } ], "created": 1751891468, "model": "deepseek-r1-250528", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 35, "completion_tokens": 159, "total_tokens": 194 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ### Go #### Request ```go package main import ( "context" "fmt" "os" "github.com/volcengine/volcengine-go-sdk/service/arkruntime" "github.com/volcengine/volcengine-go-sdk/service/arkruntime/model" "github.com/volcengine/volcengine-go-sdk/volcengine" ) func main() { client := arkruntime.NewClientWithApiKey(os.Getenv("ARK_API_KEY")) ctx := context.Background() req := model.BotChatCompletionRequest{ Model: "bot-20250707202501-k6czp", Messages: []*model.ChatCompletionMessage{ &model.ChatCompletionMessage{ Role: "system", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("You are a helpful assistant."), }, }, &model.ChatCompletionMessage{ Role: "user", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("Hello!"), }, }, }, } fmt.Println("----- standard request -----") resp, err := client.CreateBotChatCompletion(ctx, req) if err != nil { fmt.Printf("standard chat error: %v\n", err) return } if resp.References != nil { for _, ref := range resp.References { fmt.Printf("reference url: %s\n", ref.Url) } } } ``` #### Response ``` { "id": "021751891462122d1a8c9bcb9c54eb50913f5e75c28a3ac724183", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "\nHello there! How can I assist you today? 😊", "role": "assistant", "reasoning_content": "Hmm, the user just said \"Hello!\" - a simple and friendly greeting. Since this is likely the start of a conversation, they're probably testing the waters or initiating casual interaction. No complex needs here, but I should match their upbeat tone. \n\nAh, the response was straightforward - \"Hello there! How can I assist you today?\" Perfectly neutral yet warm opening. Kept it open-ended to invite follow-up. The exclamation point adds energy without being overbearing. \n\nWonder if they're new or returning? Either way, no need to overthink - just acknowledge and pivot to assistance. Short responses like this usually mean they'll either ask something specific next or vanish. Better keep the reply crisp.\n" } } ], "created": 1751891468, "model": "deepseek-r1-250528", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 35, "completion_tokens": 159, "total_tokens": 194 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ### Java #### Request ```java package com.volcengine.sample; import com.fasterxml.jackson.core.JsonProcessingException; import com.volcengine.ark.runtime.model.bot.completion.chat.BotChatCompletionRequest; import com.volcengine.ark.runtime.model.completion.chat.*; import com.volcengine.ark.runtime.model.completion.chat.ChatCompletionRequest.*; import com.volcengine.ark.runtime.service.ArkService; import java.util.ArrayList; import java.util.List; import java.util.concurrent.TimeUnit; import okhttp3.ConnectionPool; import okhttp3.Dispatcher; /* # pom.xml com.volcengine volcengine-java-sdk-ark-runtime // 替换正式版本号 LATEST */ public class Sample { static String apiKey = System.getenv("ARK_API_KEY"); static ConnectionPool connectionPool = new ConnectionPool(5, 1, TimeUnit.SECONDS); static Dispatcher dispatcher = new Dispatcher(); static ArkService service = ArkService.builder() .dispatcher(dispatcher) .connectionPool(connectionPool) .apiKey(apiKey) .baseUrl("https://ark.cn-beijing.volces.com/api/v3") .build(); public static void main(String[] args) throws JsonProcessingException { List messagesForReqList = new ArrayList<>(); ChatMessage elementForMessagesForReqList0 = ChatMessage.builder() .role(ChatMessageRole.SYSTEM) .content("You are a helpful assistant.") .build(); ChatMessage elementForMessagesForReqList1 = ChatMessage.builder().role(ChatMessageRole.USER).content("Hello!").build(); messagesForReqList.add(elementForMessagesForReqList0); messagesForReqList.add(elementForMessagesForReqList1); BotChatCompletionRequest req = BotChatCompletionRequest.builder() .model("bot-20250707202501-k6czp") .messages(messagesForReqList) .build(); service.createBotChatCompletion(req) .getChoices() .forEach(choice -> System.out.println(choice.getMessage().getContent())); // shutdown service after all requests is finished service.shutdownExecutor(); } } ``` #### Response ``` { "id": "021751891462122d1a8c9bcb9c54eb50913f5e75c28a3ac724183", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "\nHello there! How can I assist you today? 😊", "role": "assistant", "reasoning_content": "Hmm, the user just said \"Hello!\" - a simple and friendly greeting. Since this is likely the start of a conversation, they're probably testing the waters or initiating casual interaction. No complex needs here, but I should match their upbeat tone. \n\nAh, the response was straightforward - \"Hello there! How can I assist you today?\" Perfectly neutral yet warm opening. Kept it open-ended to invite follow-up. The exclamation point adds energy without being overbearing. \n\nWonder if they're new or returning? Either way, no need to overthink - just acknowledge and pivot to assistance. Short responses like this usually mean they'll either ask something specific next or vanish. Better keep the reply crisp.\n" } } ], "created": 1751891468, "model": "deepseek-r1-250528", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 35, "completion_tokens": 159, "total_tokens": 194 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ## 多轮对话 ### cURL #### Request ```bash curl https://ark.cn-beijing.volces.com/api/v3/bots/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $ARK_API_KEY" \ -d $'{ "model": "bot-20250707202501-k6czp", "messages": [ { "content": "You are a helpful assistant.", "role": "system" }, { "content": "Hello!", "role": "user" }, { "content": "Hello there! How can I assist you today? ", "role": "assistant" }, { "content": "what is your name?", "role": "user" } ] }' ``` #### Response ``` { "id": "021751892149044f7befb30d3a98790c255d78335d5360ad3db3d", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "\nGreat question! 😊 \nYou can call me **Sophie** — your friendly AI assistant. I'm here to help you with questions, ideas, tasks, or anything else you'd like to explore. How about you — what should I call you? 😊", "role": "assistant", "reasoning_content": "Hmm, the user just asked for my name after ..." } } ], "created": 1751892158, "model": "deepseek-r1-250528", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 54, "completion_tokens": 268, "total_tokens": 322 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ### Python #### Request ```python import os from volcenginesdkarkruntime import Ark # For more information，please check this document（https://ark.volcengine.com/region:cn-beijing/docs/api-key） client = Ark(api_key=os.environ.get("ARK_API_KEY")) if __name__ == "__main__": resp = client.bot_chat.completions.create( model="bot-20250707202501-k6czp", messages=[{"content":"You are a helpful assistant.","role":"system"},{"content":"Hello!","role":"user"},{"content":"Hello there! How can I assist you today? ","role":"assistant"},{"content":"what is your name?","role":"user"}], ) print(resp.choices[0].message.content) ``` #### Response ``` { "id": "021751892149044f7befb30d3a98790c255d78335d5360ad3db3d", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "\nGreat question! 😊 \nYou can call me **Sophie** — your friendly AI assistant. I'm here to help you with questions, ideas, tasks, or anything else you'd like to explore. How about you — what should I call you? 😊", "role": "assistant", "reasoning_content": "Hmm, the user just asked for my name after ..." } } ], "created": 1751892158, "model": "deepseek-r1-250528", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 54, "completion_tokens": 268, "total_tokens": 322 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ### Go #### Request ```go package main import ( "context" "fmt" "os" "github.com/volcengine/volcengine-go-sdk/service/arkruntime" "github.com/volcengine/volcengine-go-sdk/service/arkruntime/model" "github.com/volcengine/volcengine-go-sdk/volcengine" ) func main() { client := arkruntime.NewClientWithApiKey(os.Getenv("ARK_API_KEY")) ctx := context.Background() req := model.BotChatCompletionRequest{ Model: "bot-20250707202501-k6czp", Messages: []*model.ChatCompletionMessage{ &model.ChatCompletionMessage{ Role: "system", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("You are a helpful assistant."), }, }, &model.ChatCompletionMessage{ Role: "user", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("Hello!"), }, }, &model.ChatCompletionMessage{ Role: "assistant", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("Hello there! How can I assist you today? "), }, }, &model.ChatCompletionMessage{ Role: "user", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("what is your name?"), }, }, }, } fmt.Println("----- standard request -----") resp, err := client.CreateBotChatCompletion(ctx, req) if err != nil { fmt.Printf("standard chat error: %v\n", err) return } if resp.References != nil { for _, ref := range resp.References { fmt.Printf("reference url: %s\n", ref.Url) } } } ``` #### Response ``` { "id": "021751892149044f7befb30d3a98790c255d78335d5360ad3db3d", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "\nGreat question! 😊 \nYou can call me **Sophie** — your friendly AI assistant. I'm here to help you with questions, ideas, tasks, or anything else you'd like to explore. How about you — what should I call you? 😊", "role": "assistant", "reasoning_content": "Hmm, the user just asked for my name after ..." } } ], "created": 1751892158, "model": "deepseek-r1-250528", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 54, "completion_tokens": 268, "total_tokens": 322 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ### Java #### Request ```java package com.volcengine.sample; import com.fasterxml.jackson.core.JsonProcessingException; import com.volcengine.ark.runtime.model.bot.completion.chat.BotChatCompletionRequest; import com.volcengine.ark.runtime.model.completion.chat.*; import com.volcengine.ark.runtime.model.completion.chat.ChatCompletionRequest.*; import com.volcengine.ark.runtime.service.ArkService; import java.util.ArrayList; import java.util.List; import java.util.concurrent.TimeUnit; import okhttp3.ConnectionPool; import okhttp3.Dispatcher; /* # pom.xml com.volcengine volcengine-java-sdk-ark-runtime // 替换正式版本号 LATEST */ public class Sample { static String apiKey = System.getenv("ARK_API_KEY"); static ConnectionPool connectionPool = new ConnectionPool(5, 1, TimeUnit.SECONDS); static Dispatcher dispatcher = new Dispatcher(); static ArkService service = ArkService.builder() .dispatcher(dispatcher) .connectionPool(connectionPool) .apiKey(apiKey) .baseUrl("https://ark.cn-beijing.volces.com/api/v3") .build(); public static void main(String[] args) throws JsonProcessingException { List messagesForReqList = new ArrayList<>(); ChatMessage elementForMessagesForReqList0 = ChatMessage.builder() .role(ChatMessageRole.SYSTEM) .content("You are a helpful assistant.") .build(); ChatMessage elementForMessagesForReqList1 = ChatMessage.builder().role(ChatMessageRole.USER).content("Hello!").build(); ChatMessage elementForMessagesForReqList2 = ChatMessage.builder() .role(ChatMessageRole.ASSISTANT) .content("Hello there! How can I assist you today? ") .build(); ChatMessage elementForMessagesForReqList3 = ChatMessage.builder() .role(ChatMessageRole.USER) .content("what is your name?") .build(); messagesForReqList.add(elementForMessagesForReqList0); messagesForReqList.add(elementForMessagesForReqList1); messagesForReqList.add(elementForMessagesForReqList2); messagesForReqList.add(elementForMessagesForReqList3); BotChatCompletionRequest req = BotChatCompletionRequest.builder() .model("bot-20250707202501-k6czp") .messages(messagesForReqList) .build(); service.createBotChatCompletion(req) .getChoices() .forEach(choice -> System.out.println(choice.getMessage().getContent())); // shutdown service after all requests is finished service.shutdownExecutor(); } } ``` #### Response ``` { "id": "021751892149044f7befb30d3a98790c255d78335d5360ad3db3d", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "\nGreat question! 😊 \nYou can call me **Sophie** — your friendly AI assistant. I'm here to help you with questions, ideas, tasks, or anything else you'd like to explore. How about you — what should I call you? 😊", "role": "assistant", "reasoning_content": "Hmm, the user just asked for my name after ..." } } ], "created": 1751892158, "model": "deepseek-r1-250528", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 54, "completion_tokens": 268, "total_tokens": 322 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ## 流式响应 ### cURL #### Request ```bash curl https://ark.cn-beijing.volces.com/api/v3/bots/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $ARK_API_KEY" \ -d $'{ "model": "bot-20250707202501-k6czp", "messages": [ { "content": "You are a helpful assistant.", "role": "system" }, { "content": "Hello!", "role": "user" } ], "stream": true, "stream_options": { "include_usage": true } }' ``` #### Response ``` {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":"Ah"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":","},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} ... {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":".\n"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"\n","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"Hello","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} ... {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":" 😊","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant"},"finish_reason":"stop","index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","bot_usage":{"model_usage":[{"prompt_tokens":35,"completion_tokens":175,"total_tokens":210}],"action_usage":[],"action_details":[]},"metadata":{}} [DONE] ``` ### Python #### Request ```python import os from volcenginesdkarkruntime import Ark # For more information，please check this document（https://ark.volcengine.com/region:cn-beijing/docs/api-key） client = Ark(api_key=os.environ.get("ARK_API_KEY")) if __name__ == "__main__": resp = client.bot_chat.completions.create( model="bot-20250707202501-k6czp", messages=[{"content":"You are a helpful assistant.","role":"system"},{"content":"Hello!","role":"user"}], stream=True, stream_options={"include_usage":True}, ) for chunk in resp: if not chunk.choices: continue print(chunk.choices[0].delta.content, end="") ``` #### Response ``` {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":"Ah"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":","},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} ... {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":".\n"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"\n","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"Hello","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} ... {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":" 😊","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant"},"finish_reason":"stop","index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","bot_usage":{"model_usage":[{"prompt_tokens":35,"completion_tokens":175,"total_tokens":210}],"action_usage":[],"action_details":[]},"metadata":{}} [DONE] ``` ### Go #### Request ```go package main import ( "context" "fmt" "io" "os" "github.com/volcengine/volcengine-go-sdk/service/arkruntime" "github.com/volcengine/volcengine-go-sdk/service/arkruntime/model" "github.com/volcengine/volcengine-go-sdk/volcengine" ) func main() { client := arkruntime.NewClientWithApiKey(os.Getenv("ARK_API_KEY")) ctx := context.Background() req := model.BotChatCompletionRequest{ Model: "bot-20250707202501-k6czp", Messages: []*model.ChatCompletionMessage{ &model.ChatCompletionMessage{ Role: "system", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("You are a helpful assistant."), }, }, &model.ChatCompletionMessage{ Role: "user", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("Hello!"), }, }, }, Stream: true, StreamOptions: &model.StreamOptions{ IncludeUsage: true, }, } fmt.Println("----- streaming request -----") stream, err := client.CreateBotChatCompletionStream(ctx, req) if err != nil { fmt.Printf("stream chat error: %v\n", err) return } defer stream.Close() for { recv, err := stream.Recv() if err == io.EOF { return } if err != nil { fmt.Printf("Stream chat error: %v\n", err) return } if len(recv.Choices) > 0 { fmt.Print(recv.Choices[0].Delta.Content) if recv.References != nil { for _, ref := range recv.References { fmt.Printf("reference url: %s\n", ref.Url) } } } } } ``` #### Response ``` {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":"Ah"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":","},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} ... {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":".\n"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"\n","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"Hello","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} ... {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":" 😊","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant"},"finish_reason":"stop","index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","bot_usage":{"model_usage":[{"prompt_tokens":35,"completion_tokens":175,"total_tokens":210}],"action_usage":[],"action_details":[]},"metadata":{}} [DONE] ``` ### Java #### Request ```java package com.volcengine.sample; import com.fasterxml.jackson.core.JsonProcessingException; import com.volcengine.ark.runtime.model.bot.completion.chat.BotChatCompletionRequest; import com.volcengine.ark.runtime.model.completion.chat.*; import com.volcengine.ark.runtime.model.completion.chat.ChatCompletionRequest.*; import com.volcengine.ark.runtime.service.ArkService; import java.util.ArrayList; import java.util.List; import java.util.concurrent.TimeUnit; import okhttp3.ConnectionPool; import okhttp3.Dispatcher; /* # pom.xml com.volcengine volcengine-java-sdk-ark-runtime // 替换正式版本号 LATEST */ public class Sample { static String apiKey = System.getenv("ARK_API_KEY"); static ConnectionPool connectionPool = new ConnectionPool(5, 1, TimeUnit.SECONDS); static Dispatcher dispatcher = new Dispatcher(); static ArkService service = ArkService.builder() .dispatcher(dispatcher) .connectionPool(connectionPool) .apiKey(apiKey) .baseUrl("https://ark.cn-beijing.volces.com/api/v3") .build(); public static void main(String[] args) throws JsonProcessingException { List messagesForReqList = new ArrayList<>(); ChatMessage elementForMessagesForReqList0 = ChatMessage.builder() .role(ChatMessageRole.SYSTEM) .content("You are a helpful assistant.") .build(); ChatMessage elementForMessagesForReqList1 = ChatMessage.builder().role(ChatMessageRole.USER).content("Hello!").build(); messagesForReqList.add(elementForMessagesForReqList0); messagesForReqList.add(elementForMessagesForReqList1); ChatCompletionRequestStreamOptions streamOptionsForReq = new ChatCompletionRequestStreamOptions(true); BotChatCompletionRequest req = BotChatCompletionRequest.builder() .model("bot-20250707202501-k6czp") .messages(messagesForReqList) .stream(true) .streamOptions(streamOptionsForReq) .build(); service.streamBotChatCompletion(req) .doOnError(Throwable::printStackTrace) .blockingForEach( choice -> { if (choice.getChoices().size() > 0) { System.out.print( choice.getChoices().get(0).getMessage().getContent()); } }); // shutdown service after all requests is finished service.shutdownExecutor(); } } ``` #### Response ``` {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":"Ah"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":","},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} ... {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant","reasoning_content":".\n"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"\n","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"Hello","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} ... {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":" 😊","role":"assistant"},"index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[{"delta":{"content":"","role":"assistant"},"finish_reason":"stop","index":0}],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","metadata":{}} {"id":"0217518925181918b15924cfe8caf3be8175ae89769e0203c419c","choices":[],"created":1751892518,"model":"deepseek-r1-250528","object":"chat.completion.chunk","bot_usage":{"model_usage":[{"prompt_tokens":35,"completion_tokens":175,"total_tokens":210}],"action_usage":[],"action_details":[]},"metadata":{}} [DONE] ``` ## 群聊 ### cURL #### Request ```bash curl https://ark.cn-beijing.volces.com/api/v3/bots/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $ARK_API_KEY" \ -d $'{ "model": "bot-20250707205612-rr7gm", "messages": [ { "content": "你是谁？", "role": "user" } ], "metadata": { "group_chat_config": { "characters": [ { "model_desc": { "endpoint_id": "ep-20250703110544-ht42m" }, "name": "孙悟空", "system_prompt": "齐天大圣孙悟空，生性聪明、活泼、忠诚、嫉恶如仇，有责任心，偶尔也有些骄傲自负。" } ], "description": "遇到妖精", "user_name": "唐僧" }, "target_character_name": "孙悟空" }, "stream": false, "stream_options": { "include_usage": false } }' ``` #### Response ``` { "id": "021751893662764bd5b399bfe40c748dcd249c43e3efca02ca61d", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "用户现在需要我以孙悟空的身份回复唐僧的询问。首先要体现出孙悟空的性格，语气要有点傲娇又带着不耐烦但又带着忠诚的感觉。那应该这样回复：“俺乃齐天大圣孙悟空！你这呆子，怎地连俺都认不得了？俺可是要护你去西天取经的啊！”俺乃齐天大圣孙悟空！你这秃驴莫不是糊涂了？俺老孙在此，正是要保你去西天取经呢，还能是谁？", "role": "assistant", "reasoning_content": "俺老孙乃是齐天大圣孙悟空！你这肉眼凡胎的和尚，连俺都认不出来啦？俺可是保护你去西天取经的大徒弟啊！" } } ], "created": 1751893664, "model": "doubao-seed-1-6-flash-250615", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 149, "completion_tokens": 143, "total_tokens": 292 } ], "action_usage": [], "action_details": [] }, "metadata": { "group_chat_config": { "user_name": "唐僧", "description": "遇到妖精", "characters": [ { "name": "孙悟空", "system_prompt": "齐天大圣孙悟空，生性聪明、活泼、忠诚、嫉恶如仇，有责任心，偶尔也有些骄傲自负。", "model_desc": { "endpoint_id": "ep-20250703110544-ht42m" } } ] }, "target_character_name": "孙悟空" } } ``` ### Python #### Request ```python import os from volcenginesdkarkruntime import Ark # For more information，please check this document（https://ark.volcengine.com/region:cn-beijing/docs/api-key） client = Ark(api_key=os.environ.get("ARK_API_KEY")) if __name__ == "__main__": resp = client.bot_chat.completions.create( model="bot-20250707205612-rr7gm", messages=[{"content":"你是谁？","role":"user"}], stream=False, stream_options={"include_usage":False}, metadata={"group_chat_config":{"characters":[{"model_desc":{"endpoint_id":"ep-20250703110544-ht42m"},"name":"孙悟空","system_prompt":"齐天大圣孙悟空，生性聪明、活泼、忠诚、嫉恶如仇，有责任心，偶尔也有些骄傲自负。"}],"description":"遇到妖精","user_name":"唐僧"},"target_character_name":"孙悟空"}, ) print(resp.choices[0].message.content) ``` #### Response ``` { "id": "021751893662764bd5b399bfe40c748dcd249c43e3efca02ca61d", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "用户现在需要我以孙悟空的身份回复唐僧的询问。首先要体现出孙悟空的性格，语气要有点傲娇又带着不耐烦但又带着忠诚的感觉。那应该这样回复：“俺乃齐天大圣孙悟空！你这呆子，怎地连俺都认不得了？俺可是要护你去西天取经的啊！”俺乃齐天大圣孙悟空！你这秃驴莫不是糊涂了？俺老孙在此，正是要保你去西天取经呢，还能是谁？", "role": "assistant", "reasoning_content": "俺老孙乃是齐天大圣孙悟空！你这肉眼凡胎的和尚，连俺都认不出来啦？俺可是保护你去西天取经的大徒弟啊！" } } ], "created": 1751893664, "model": "doubao-seed-1-6-flash-250615", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 149, "completion_tokens": 143, "total_tokens": 292 } ], "action_usage": [], "action_details": [] }, "metadata": { "group_chat_config": { "user_name": "唐僧", "description": "遇到妖精", "characters": [ { "name": "孙悟空", "system_prompt": "齐天大圣孙悟空，生性聪明、活泼、忠诚、嫉恶如仇，有责任心，偶尔也有些骄傲自负。", "model_desc": { "endpoint_id": "ep-20250703110544-ht42m" } } ] }, "target_character_name": "孙悟空" } } ``` ### Go #### Request ```go package main import ( "context" "fmt" "os" "github.com/volcengine/volcengine-go-sdk/service/arkruntime" "github.com/volcengine/volcengine-go-sdk/service/arkruntime/model" "github.com/volcengine/volcengine-go-sdk/volcengine" ) func main() { client := arkruntime.NewClientWithApiKey(os.Getenv("ARK_API_KEY")) ctx := context.Background() req := model.BotChatCompletionRequest{ Model: "bot-20250707205612-rr7gm", Messages: []*model.ChatCompletionMessage{ &model.ChatCompletionMessage{ Role: "user", Content: &model.ChatCompletionMessageContent{ StringValue: volcengine.String("你是谁？"), }, }, }, Stream: false, StreamOptions: &model.StreamOptions{ IncludeUsage: false, }, Metadata: map[string]interface{}{ "group_chat_config": map[string]interface{}{"characters": []interface{}{map[string]interface{}{"model_desc": map[string]interface{}{"endpoint_id": "ep-20250703110544-ht42m"}, "name": "孙悟空", "system_prompt": "齐天大圣孙悟空，生性聪明、活泼、忠诚、嫉恶如仇，有责任心，偶尔也有些骄傲自负。"}}, "description": "遇到妖精", "user_name": "唐僧"}, "target_character_name": "孙悟空", }, } fmt.Println("----- standard request -----") resp, err := client.CreateBotChatCompletion(ctx, req) if err != nil { fmt.Printf("standard chat error: %v\n", err) return } fmt.Println(*resp.Choices[0].Message.Content.StringValue) if resp.References != nil { for _, ref := range resp.References { fmt.Printf("reference url: %s\n", ref.Url) } } } ``` #### Response ``` { "id": "021751893662764bd5b399bfe40c748dcd249c43e3efca02ca61d", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "用户现在需要我以孙悟空的身份回复唐僧的询问。首先要体现出孙悟空的性格，语气要有点傲娇又带着不耐烦但又带着忠诚的感觉。那应该这样回复：“俺乃齐天大圣孙悟空！你这呆子，怎地连俺都认不得了？俺可是要护你去西天取经的啊！”俺乃齐天大圣孙悟空！你这秃驴莫不是糊涂了？俺老孙在此，正是要保你去西天取经呢，还能是谁？", "role": "assistant", "reasoning_content": "俺老孙乃是齐天大圣孙悟空！你这肉眼凡胎的和尚，连俺都认不出来啦？俺可是保护你去西天取经的大徒弟啊！" } } ], "created": 1751893664, "model": "doubao-seed-1-6-flash-250615", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 149, "completion_tokens": 143, "total_tokens": 292 } ], "action_usage": [], "action_details": [] }, "metadata": { "group_chat_config": { "user_name": "唐僧", "description": "遇到妖精", "characters": [ { "name": "孙悟空", "system_prompt": "齐天大圣孙悟空，生性聪明、活泼、忠诚、嫉恶如仇，有责任心，偶尔也有些骄傲自负。", "model_desc": { "endpoint_id": "ep-20250703110544-ht42m" } } ] }, "target_character_name": "孙悟空" } } ``` ### Java #### Request ```java package com.volcengine.sample; import com.fasterxml.jackson.core.JsonProcessingException; import com.fasterxml.jackson.databind.ObjectMapper; import com.volcengine.ark.runtime.model.bot.completion.chat.BotChatCompletionRequest; import com.volcengine.ark.runtime.model.completion.chat.*; import com.volcengine.ark.runtime.model.completion.chat.ChatCompletionRequest.*; import com.volcengine.ark.runtime.service.ArkService; import java.util.ArrayList; import java.util.List; import java.util.Map; import java.util.concurrent.TimeUnit; import okhttp3.ConnectionPool; import okhttp3.Dispatcher; /* # pom.xml com.volcengine volcengine-java-sdk-ark-runtime // 替换正式版本号 LATEST */ public class Sample { static String apiKey = System.getenv("ARK_API_KEY"); static ConnectionPool connectionPool = new ConnectionPool(5, 1, TimeUnit.SECONDS); static Dispatcher dispatcher = new Dispatcher(); static ArkService service = ArkService.builder() .dispatcher(dispatcher) .connectionPool(connectionPool) .apiKey(apiKey) .baseUrl("https://ark.cn-beijing.volces.com/api/v3") .build(); public static void main(String[] args) throws JsonProcessingException { List messagesForReqList = new ArrayList<>(); ChatMessage elementForMessagesForReqList0 = ChatMessage.builder().role(ChatMessageRole.USER).content("你是谁？").build(); messagesForReqList.add(elementForMessagesForReqList0); ChatCompletionRequestStreamOptions streamOptionsForReq = new ChatCompletionRequestStreamOptions(false); BotChatCompletionRequest req = BotChatCompletionRequest.builder() .model("bot-20250707205612-rr7gm") .messages(messagesForReqList) .metadata( (convertMap( "{\"group_chat_config\":{\"characters\":[{\"model_desc\":{\"endpoint_id\":\"ep-20250703110544-ht42m\"},\"name\":\"孙悟空\",\"system_prompt\":\"齐天大圣孙悟空，生性聪明、活泼、忠诚、嫉恶如仇，有责任心，偶尔也有些骄傲自负。\"}],\"description\":\"遇到妖精\",\"user_name\":\"唐僧\"},\"target_character_name\":\"孙悟空\"}"))) .build(); service.createBotChatCompletion(req) .getChoices() .forEach(choice -> System.out.println(choice.getMessage().getContent())); // shutdown service after all requests is finished service.shutdownExecutor(); } public static Map convertMap(String jsonString) throws JsonProcessingException { ObjectMapper objectMapper = new ObjectMapper(); Map map = objectMapper.readValue(jsonString, Map.class); return map; } } ``` #### Response ``` { "id": "021751893662764bd5b399bfe40c748dcd249c43e3efca02ca61d", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "用户现在需要我以孙悟空的身份回复唐僧的询问。首先要体现出孙悟空的性格，语气要有点傲娇又带着不耐烦但又带着忠诚的感觉。那应该这样回复：“俺乃齐天大圣孙悟空！你这呆子，怎地连俺都认不得了？俺可是要护你去西天取经的啊！”俺乃齐天大圣孙悟空！你这秃驴莫不是糊涂了？俺老孙在此，正是要保你去西天取经呢，还能是谁？", "role": "assistant", "reasoning_content": "俺老孙乃是齐天大圣孙悟空！你这肉眼凡胎的和尚，连俺都认不出来啦？俺可是保护你去西天取经的大徒弟啊！" } } ], "created": 1751893664, "model": "doubao-seed-1-6-flash-250615", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 149, "completion_tokens": 143, "total_tokens": 292 } ], "action_usage": [], "action_details": [] }, "metadata": { "group_chat_config": { "user_name": "唐僧", "description": "遇到妖精", "characters": [ { "name": "孙悟空", "system_prompt": "齐天大圣孙悟空，生性聪明、活泼、忠诚、嫉恶如仇，有责任心，偶尔也有些骄傲自负。", "model_desc": { "endpoint_id": "ep-20250703110544-ht42m" } } ] }, "target_character_name": "孙悟空" } } ``` ## 视觉理解 ### cURL #### Request ```bash curl https://ark.cn-beijing.volces.com/api/v3/bots/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $ARK_API_KEY" \ -d $'{ "model": "bot-20250707202501-k6czp", "messages": [ { "content": [ { "image_url": { "url": "https://arkdocs.tos-cn-beijing.volces.com/images/advanced-usage/ark_demo_img_1-zh.png" }, "type": "image_url" }, { "text": "支持输入是图片的模型系列是哪个？", "type": "text" } ], "role": "user" } ] }' ``` #### Response ``` { "id": "021751894443607bd1685776f26233986d9c881482ac5cc63c00a", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "支持输入是图片的模型系列是**Doubao-1.5-vision**。\n\n在表格中，“输入”列的“图像”项中，仅Doubao-1.5-vision对应的单元格标记为“√”，表示该模型系列支持图像输入。", "role": "assistant", "reasoning_content": "\n我现在需要解决的问题是：“支持输入是图片的模型系列是哪个？..." } } ], "created": 1751894450, "model": "doubao-seed-1-6-250615", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 1131, "completion_tokens": 310, "total_tokens": 1441 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ### Python #### Request ```python import os from volcenginesdkarkruntime import Ark # For more information，please check this document（https://ark.volcengine.com/region:cn-beijing/docs/api-key） client = Ark(api_key=os.environ.get("ARK_API_KEY")) if __name__ == "__main__": resp = client.bot_chat.completions.create( model="bot-20250707202501-k6czp", messages=[{"content":[{"image_url":{"url":"https://arkdocs.tos-cn-beijing.volces.com/images/advanced-usage/ark_demo_img_1-zh.png"},"type":"image_url"},{"text":"支持输入是图片的模型系列是哪个？","type":"text"}],"role":"user"}], ) print(resp.choices[0].message.content) ``` #### Response ``` { "id": "021751894443607bd1685776f26233986d9c881482ac5cc63c00a", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "支持输入是图片的模型系列是**Doubao-1.5-vision**。\n\n在表格中，“输入”列的“图像”项中，仅Doubao-1.5-vision对应的单元格标记为“√”，表示该模型系列支持图像输入。", "role": "assistant", "reasoning_content": "\n我现在需要解决的问题是：“支持输入是图片的模型系列是哪个？..." } } ], "created": 1751894450, "model": "doubao-seed-1-6-250615", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 1131, "completion_tokens": 310, "total_tokens": 1441 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ### Go #### Request ```go package main import ( "context" "fmt" "os" "github.com/volcengine/volcengine-go-sdk/service/arkruntime" "github.com/volcengine/volcengine-go-sdk/service/arkruntime/model" ) func main() { client := arkruntime.NewClientWithApiKey(os.Getenv("ARK_API_KEY")) ctx := context.Background() req := model.BotChatCompletionRequest{ Model: "bot-20250707202501-k6czp", Messages: []*model.ChatCompletionMessage{ &model.ChatCompletionMessage{ Role: "user", Content: &model.ChatCompletionMessageContent{ ListValue: []*model.ChatCompletionMessageContentPart{ &model.ChatCompletionMessageContentPart{ Type: model.ChatCompletionMessageContentPartTypeImageURL, ImageURL: &model.ChatMessageImageURL{ URL: "https://arkdocs.tos-cn-beijing.volces.com/images/advanced-usage/ark_demo_img_1-zh.png", }, }, &model.ChatCompletionMessageContentPart{ Type: model.ChatCompletionMessageContentPartTypeText, Text: "支持输入是图片的模型系列是哪个？", }, }, }, }, }, } fmt.Println("----- standard request -----") resp, err := client.CreateBotChatCompletion(ctx, req) if err != nil { fmt.Printf("standard chat error: %v\n", err) return } if resp.References != nil { for _, ref := range resp.References { fmt.Printf("reference url: %s\n", ref.Url) } } } ``` #### Response ``` { "id": "021751894443607bd1685776f26233986d9c881482ac5cc63c00a", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "支持输入是图片的模型系列是**Doubao-1.5-vision**。\n\n在表格中，“输入”列的“图像”项中，仅Doubao-1.5-vision对应的单元格标记为“√”，表示该模型系列支持图像输入。", "role": "assistant", "reasoning_content": "\n我现在需要解决的问题是：“支持输入是图片的模型系列是哪个？..." } } ], "created": 1751894450, "model": "doubao-seed-1-6-250615", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 1131, "completion_tokens": 310, "total_tokens": 1441 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ``` ### Java #### Request ```java package com.volcengine.sample; import com.fasterxml.jackson.core.JsonProcessingException; import com.volcengine.ark.runtime.model.bot.completion.chat.BotChatCompletionRequest; import com.volcengine.ark.runtime.model.completion.chat.*; import com.volcengine.ark.runtime.model.completion.chat.ChatCompletionRequest.*; import com.volcengine.ark.runtime.service.ArkService; import java.util.ArrayList; import java.util.List; import java.util.concurrent.TimeUnit; import okhttp3.ConnectionPool; import okhttp3.Dispatcher; /* # pom.xml com.volcengine volcengine-java-sdk-ark-runtime // 替换正式版本号 LATEST */ public class Sample { static String apiKey = System.getenv("ARK_API_KEY"); static ConnectionPool connectionPool = new ConnectionPool(5, 1, TimeUnit.SECONDS); static Dispatcher dispatcher = new Dispatcher(); static ArkService service = ArkService.builder() .dispatcher(dispatcher) .connectionPool(connectionPool) .apiKey(apiKey) .baseUrl("https://ark.cn-beijing.volces.com/api/v3") .build(); public static void main(String[] args) throws JsonProcessingException { List messagesForReqList = new ArrayList<>(); List contentForElementForMessagesForReqList0 = new ArrayList<>(); // 创建并设置图片内容部分 ChatCompletionContentPart elementForContentForElementForMessagesForReqList00 = ChatCompletionContentPart.builder() .type("image_url") .imageUrl(new ChatCompletionContentPart.ChatCompletionContentPartImageURL( "https://arkdocs.tos-cn-beijing.volces.com/images/advanced-usage/ark_demo_img_1-zh.png")) .build(); ChatCompletionContentPart elementForContentForElementForMessagesForReqList01 = ChatCompletionContentPart.builder() .type("text") .text("支持输入是图片的模型系列是哪个？") .build(); contentForElementForMessagesForReqList0.add( elementForContentForElementForMessagesForReqList00); contentForElementForMessagesForReqList0.add( elementForContentForElementForMessagesForReqList01); // 创建用户消息对象并设置内容 ChatMessage elementForMessagesForReqList0 = ChatMessage.builder() .role(ChatMessageRole.USER) .multiContent(contentForElementForMessagesForReqList0) .build(); // 添加用户消息到消息列表中 messagesForReqList.add(elementForMessagesForReqList0); BotChatCompletionRequest req = BotChatCompletionRequest.builder() .model("bot-20250707202501-k6czp") .messages(messagesForReqList) .build(); service.createBotChatCompletion(req) .getChoices() .forEach(choice -> System.out.println(choice.getMessage().getContent())); // shutdown service after all requests is finished service.shutdownExecutor(); } } ``` #### Response ``` { "id": "021751894443607bd1685776f26233986d9c881482ac5cc63c00a", "choices": [ { "finish_reason": "stop", "index": 0, "message": { "content": "支持输入是图片的模型系列是**Doubao-1.5-vision**。\n\n在表格中，“输入”列的“图像”项中，仅Doubao-1.5-vision对应的单元格标记为“√”，表示该模型系列支持图像输入。", "role": "assistant", "reasoning_content": "\n我现在需要解决的问题是：“支持输入是图片的模型系列是哪个？..." } } ], "created": 1751894450, "model": "doubao-seed-1-6-250615", "object": "chat.completion", "bot_usage": { "model_usage": [ { "prompt_tokens": 1131, "completion_tokens": 310, "total_tokens": 1441 } ], "action_usage": [], "action_details": [] }, "metadata": {} } ```