跳到主要内容
AI 视频模型
Kling AI
Kling AI Avatar

Kling AI Avatar Standard

用 1 张图片和 1 段目标音频尝试制作说话头像视频,提示词可以留空。可先用短音频检查口型、停顿和人物表现,再决定是否采用结果。

素材
单图+目标音频
预扣
每音频秒 8 积分
提示词
可留空;上限 5,000 字符
输入音频
最长 5 分钟,先完成上传
快速了解
  • 准备图片与独立目标音频即可开始,不必先写提示词
  • 按核实的音频时长估算预扣,结算另有条件
  • 图片格式、音频时长和素材校验信息均需满足要求
  • 先做短片段试样,再检查口型、表情和停顿

相关查询: Kling AI Avatar Standard 怎么用 · Kling AI Avatar Standard 如何预扣积分 · Kling AI Avatar Standard 需要什么素材

示例

Cinematic scene

"Create a 15-second realistic cinematic slice-of-life video in 16:9, 24fps. Visual-only video, no dialogue, no voice, no music, no sound effects, no subtitles, no text on screen. Warm golden-hour sunlight, soft handheld camera, muted film color grade, slight grain, natural shadows, peaceful nostalgic neighborhood mood. Fully realistic live-action style, not game-like, not animated. Main character: a young Korean woman in her early 20s with messy dark brown hair tied in a loose ponytail, soft natural face, calm gentle expression, dark gray cropped sleeveless tank top, loose light-wash blue jeans, black necklace, casual sneakers. Keep her look consistent in every shot. Location: a long narrow real-life residential alley between tall tan stone and concrete walls. Sandy dusty ground, warm sunlight glowing on one side, soft tree shadows across the path, palm trees and green leaves visible above the walls, a blue painted wall far ahead, one bicycle leaning near the end of the alley, quiet empty neighborhood feeling. 0:00–0:02: Rear tracking shot. The Korean woman walks slowly through the long narrow alley from behind. Golden sunlight stretches across the tan walls. Camera follows with natural handheld motion. 0:02–0:04: Wide alley shot. A small gray-brown tabby cat appears from a low gap near the left side wall and steps into the alley. The woman notices it, slows down, and turns slightly toward it. 0:04–0:06: Medium side shot. The cat walks near her feet and looks up. She smiles softly, bends down slowly, and extends one hand in a gentle careful way. 0:06–0:08: Low close-up. The cat sniffs her fingers, then rubs its head against her hand. She gently pets the cat’s head and back. Show realistic fur texture, tiny whisker movement, and natural cat behavior. 0:08–0:10: Emotional close-up. The woman smiles warmly while petting the cat. Sunlight flickers through tree leaves across her face, shoulders, and hair. Keep the moment soft, calm, and human. 0:10–0:12: Medium-wide shot. The cat walks ahead a few steps, then looks back at her. She stands slowly and follows it with a curious peaceful smile. 0:12–0:14: Front tracking shot. She walks toward camera with the cat beside her near the wall. She looks down at the cat and smiles. The long alley, tan walls, blue wall far behind, trees, and bicycle remain visible. 0:14–0:15: Final close shot. The cat brushes against her leg. She looks toward the camera with a soft peaceful smile, then looks down at the cat again. Hold the warm golden alley atmosphere for a gentle cinematic ending."

Cinematic scene

30-second photorealistic cinematic Western black comedy on a dusty American ranch. Strictly preserve all reference identities, clothing, proportions, animals, props, environment, and visual style. Style: gritty 1970s Western, 35mm grain, faded warm colors, dusty sunlight, high contrast, film scratches, lens flare, anamorphic framing, shallow DOF, realistic textures. Serious Western cinematography contrasting with absurd physical comedy. Never cartoonish. 0–2s: Side-angle close-up of the cowboy’s hand confidently gripping the handheld prop from the reference. Show its full recognizable silhouette, metallic sunlight highlight, drifting dust. 2–7s: Wide ranch shot. Cowboy confidently spins the prop twice, catches it, adjusts his hat, smirks and says: “Damn, I look good.” Walks forward like the coolest man alive. 7–13s: Side view showing cowboy approaching the milk worker and dairy cow together. Cowboy: “Need a hand?” Milk worker: “You know how to do this?” Cowboy: “Of course.” He sits beside the cow, confidently positions his hands to milk it while the worker watches skeptically. 13–20s: Worker begins: “First, you just…” Cowboy interrupts: “I got it.” He milks incorrectly and squeezes too hard. Milk shoots into his face. CUT to front-facing slow motion as milk covers his face, hair, hat, shirt, vest and shoulders. He freezes in stunned disbelief. Silence except dripping milk. 20–26s: Slow push-in. Cowboy remains completely emotionless and defeated, milk dripping everywhere. Worker tries not to laugh, fails and bursts out laughing. Cowboy ignores him and stares into camera. 26–30s: Extreme close-up of the cow staring judgmentally into camera. It slowly blinks, then deadpan: “Moooo.” Hard cut to black with vintage film flicker and projector sound. No subtitles, captions, text or logos. Natural synchronized dialogue. Cinematic intentional camera movement. Start as a serious stylish Western; gradually become absurd deadpan comedy. Keep all humans, animals, props and environments photorealistic.

Cinematic scene

Main Subject: Young American woman, early 20s, natural everyday appearance, cropped puffer vest over a long-sleeve tee, biker shorts, white sneakers, hair in a high ponytail, minimal makeup, realistic skin texture, energetic morning-person personality. Location: Small neighborhood street with a corner coffee shop, early morning. Brick sidewalks, parked bicycles, a bench outside, soft morning fog lifting. No large crowds, no chain branding. Visual Style: Ultra-realistic documentary realism, candid unscripted feeling, cool soft morning light. Camera Style: Early 2000s consumer DV camcorder aesthetic, handheld shake, autofocus hunting between her and the shop window, faded soft contrast, mild sensor noise. No stabilization. Timeline (15 sec, each slot = 2 compressed beats): 00:00–00:03 → She jogs up to the coffee shop, then pauses to catch her breath. 00:03–00:06 → She turns to camera saying "okay I definitely earned this coffee," laughing. 00:06–00:09 → She steps inside briefly, then walks back out holding a cup. 00:09–00:12 → She takes a sip, then sighs happily, warming her hands on the cup. 00:12–00:15 → She looks at camera saying "best part of my morning, no joke," smiling as it fades. Audio: Distant footsteps, faint café chatter, birds, light wind. Her dialogue as noted above. No music. Goal: A warm, energetic morning routine moment relatable, cheerful, real.

Cinematic scene

One original female character with identity perfectly locked across the entire video. The same face, hairstyle, body shape, clothing, and accessories remain completely consistent. She has short chestnut hair tucked behind one ear, wears a mustard-yellow oversized hoodie, black athletic shorts, white crew socks, vintage running shoes, and a simple digital wristwatch. Location: a quiet countryside railway crossing at sunrise. Empty train tracks stretching into the distance, open grassy fields, wildflowers, old wooden fences, a small railway signal, morning mist, scattered birds, and distant hills. No houses, no streets, no vehicles, no city elements, no shops, and no crowds. The camera feels like an old consumer camcorder from the early 2000s. Unpredictable handheld movement, accidental zooming, exposure constantly adjusting, delayed autofocus, washed-out colors, visible digital noise, tape compression artifacts, soft image quality, slight motion blur, and completely unstabilized footage. It should feel like an amateur family recording, never cinematic. She begins sitting on the railway platform edge, tying her shoelaces while looking toward the empty tracks. She stands and slowly walks through tall grass, brushing wildflowers with her fingertips. A gust of wind lifts her hoodie as she laughs naturally. She discovers a paper airplane on the ground, unfolds it, smiles with curiosity, and gently tosses it into the air. The camera struggles to keep focus as it drifts away. She then stands beside the crossing signal, closes her eyes for a brief moment, taking in the morning breeze. In the final shot she walks away across the open field, turns once with a bright genuine smile, gives a small wave without speaking, and the recording suddenly stops mid-motion as if the camcorder battery died. Natural audio only: birds singing, insects, distant wind across the fields, grass rustling, soft footsteps, occasional railway signal clicking, and ambient countryside sounds. No music, narration, artificial effects, or post-production audio.

Fantasy scene

Fantasy scene

一句话了解

这个模型怎么样?

先准备 JPEG/PNG 图片,文件不超过 10 MiB。目标音频支持 MPEG、WAV、X-WAV、AAC、MP4 或 OGG;本站上传的音频文件不超过 95 MiB,输入时长不超过 5 分钟。文件较大时可先适当压缩,确认能完成上传,再提交任务;上传后仍需通过素材和音频时长检查。

模型定位
以单图和目标音频为素材,尝试制作说话头像视频。
适合谁用
可用于尝试课程讲解、商品说明或简短通知;人物、语言、口型和动作是否合适,需要用自己的素材检查。
本站使用说明
准备好素材后,先查看音频时长对应的预扣估算,再确认是否提交。生成后检查结果,并及时下载需要保留的文件。

关键参数一览

根据这些信息,判断模型是否适合你的任务。

模型类别
AI 视频模型
厂商
Kling AI
模型家族
Kling AI Avatar
运行方式
异步任务
图片要求
JPEG/PNG,不超过 10 MiB
音频准备
本站音频上传不超过 95 MiB,输入时长不超过 5 分钟
可调选项
暂无其他可调选项
结算说明
成功后仍需确认有效状态与实际消耗;异常保留预扣待核

它还叫什么

同一个模型在公开资料中可能有不同名称,这里整理相关称呼,方便查找和比较。

Kling AI Avatar Standard

用户常见问题

从任务、输入条件和结果要求出发,整理选择模型时需要确认的问题。

Kling AI Avatar Standard 怎么用
Kling AI Avatar Standard 如何预扣积分
Kling AI Avatar Standard 需要什么素材
Kling AI Avatar 提示词可以留空吗
Kling AI Avatar 输入音频最长多久
图片和音频怎样做数字人口播
Kling AI Avatar Standard 与 Pro 有什么不同
Kling AI Avatar Standard 与 Infinitalk 对比
课程讲解音频怎样准备
商品口播试样要检查什么

选型指南

根据输入素材、任务要求和积分成本,逐项确认模型是否适合。

1
已有图片和配音时,先做短片段试样

选择清楚且有权使用的图片与音频,先检查口型、停顿和身份一致性。短片段合适,再安排更长内容;每段输入音频仍须满足时长要求。

2
比较 Standard 与 Pro 时,分开看输入和费用

两档都需要图片和音频,提示词可留空;本站预扣系数不同。不要只凭档位名称判断成片清晰度或效果,仍需查看各自当前要求并检查结果。

3
需要调分辨率或 seed 时,再核对 Infinitalk

Standard 暂无其他可调选项。Infinitalk 提供分辨率与 seed 选项,但需要非空提示词;选择前还应核对它自己的素材、时长与费用要求。

模型对比

把常见候选放在同一任务下比较,帮助你确认输入方式、控制项和成本差异。

Standard 和 Pro 的素材要求是否相同?
比较两档费用时,应先看什么?
Standard 与 Infinitalk 对提示词有什么不同?
哪些模型提供可调分辨率或 seed?

模型横向对比表

按同一组维度比较当前模型和候选模型的输入、设置与使用边界。

素材准备
Kling AI Avatar Standard
1 张 JPEG/PNG 图片+1 段目标音频
Kling AI Avatar Pro
1 张 JPEG/PNG 图片+1 段目标音频
Infinitalk
1 张图片+1 段目标音频
提示词
Kling AI Avatar Standard
可以留空;本站长度上限 5,000 字符
Kling AI Avatar Pro
可以留空
Infinitalk
必须填写非空提示词
预扣估算
Kling AI Avatar Standard
每音频秒 8 积分,合计向上取整
Kling AI Avatar Pro
每音频秒 16 积分,合计向上取整
Infinitalk
按核实的音频时长与所选分辨率估算
可调选项
Kling AI Avatar Standard
暂无其他可调选项
Kling AI Avatar Pro
暂无其他可调选项
Infinitalk
可选 480p/720p 请求分辨率,并提供 seed 选项

适用情况

依据公开资料与当前站内设置,说明模型适用的任务及需要注意的限制。

先确认目标音频,再安排口播内容

预扣估算依据本站核实的目标音频时长。压缩和剪辑后,请重新确认音频内容、停顿和上传状态,再提交任务。

换配音后,重新检查人物表现

同一张图片换用其他内容或语言的音频,仍需检查口型、表情、动作与语义是否协调,不能把前一次结果当作所有素材的效果保证。

能力详解

配合图片与音频制作说话头像

使用单图与独立目标音频尝试生成口播视频。人物是否稳定、嘴部动作是否自然,需要逐段看结果。

提示词可省略

不填写提示词也可准备提交;若要补充画面意图,可写简短说明。文字不能代替必需的图片与音频,也不保证口型或动作效果。

按音频长度查看预扣估算

准备好音频后,可查看对应的费用估算;提交时仍会核实时长。最终结算还取决于任务状态和有效实际消耗。

适合做什么

课程与知识讲解

用课程音频和授权人物图试做讲解片段,检查口型、停顿与身份一致性,再决定能否采用。

商品口播试样

先用简短商品音频制作试样,核对人物表现、表达节奏与信息是否准确,再决定是否继续制作。

通知与培训片段

可尝试用授权素材制作简短说明。较长配音先拆成多个片段,每段输入音频不超过 5 分钟,并分别检查结果及衔接。

提示词技巧

提示词只补充必要的画面意图

提示词可以留空。需要填写时,用简短说明表达希望的呈现,避免把整段音频内容或素材格式清单重复写入;结果仍需检查。

先听一遍目标音频

尽量减少噪声,确认说话内容、停顿和音量适合口播。压缩文件后再听一遍,避免为了上传而漏掉内容或影响听感。

检查图片清晰度与使用授权

确认图片内容清楚且有权使用,再与目标音频配合试做。面部、嘴部、表情与身份一致性都应在生成后检查。

为什么选它

已有图片和目标音频即可开始准备,提示词可以留空。
提交前能按音频时长查看预扣估算。
暂无其他可调选项,准备重点放在素材和音频。

使用前要知道

口型、表情、停顿、身份一致性与不同语言的表现都需要人工检查。
输入音频不超过 5 分钟;分段制作后仍需检查每段结果与衔接。
本站音频上传不超过 95 MiB,符合大小仍需通过素材校验;需要保存的结果请及时下载备份。

可调参数

根据这些信息,判断模型是否适合你的任务。

当前页面未列出额外可调设置,请以工作台为准。

积分消耗

每音频秒预扣 8 积分,合计向上取整

提交前可查看费用估算。本站核实目标音频时长后,按 8 积分乘以音频秒数计算合计并向上取整,任务受理后预扣。任务成功后,还需确认最终状态与有效实际消耗;只有向上取整的实际消耗不超过预扣,才会结算并退还差额。用量缺失、无效,查询或状态异常,或实际消耗超过预扣时,继续保留预扣待核,不自动追加扣点。失败任务请查看对应扣退明细。

使用建议

准备多次生成时,建议先用少量素材测试,核对输入设置和结果是否适合任务,再决定是否继续。每次生成均按工作台显示的计费规则扣点。

常见问题

相关模型

按你的任务比较这些模型的输入要求、功能和当前状态,再查看对应详情。

Kling AI Avatar Pro

用人物图配合已有音频试做口播,先以品牌试样或课程片段检查口型、动作与画面,再决定是否用于交付。

AI 视频模型
本站预扣:16 积分 / 音频秒

Infinitalk

先把要说的话整理成短音频,再配人物图片与非空提示词试做口播。检查人物、语言、口型、动作与构图后,再决定是否用于课程说明、客服答复或虚拟主播内容。

AI 视频模型
本站预扣:480p 每音频秒 3 积分;720p 每音频秒 12 积分

Seedance 2.5

用 Seedance 2.5 规划较长镜头,先选择文字、首帧与首尾帧或多模态参考场景。请求时长可设为最长 30 秒;默认 5 秒,选择自动时长则按 30 秒预留预计积分。准备素材时,数量、单段时长与累计时长需分别核对。

AI 视频模型
预计预扣 102–5,700 积分

信息来源

页面内容依据公开模型资料与 Kyeo AI 的模型信息整理。能否在线使用,请以当前模型状态为准。

来源说明

本页结合本型号的归档接口资料与本站规则,说明图音素材、提示词、费用和结果保存。归档资料不代替实际成片检查。

内容复核日期: 2026-08-28
Kling AI Avatar Standard 接口资料(归档)
平台

用于核对本型号的图音输入与提示词说明;本站上传、预扣、结算和结果保存按本站规则处理,归档资料不代替实际成片检查。